tuprints.ulb.tu-darmstadt.de Gao... · A Dissertation approved by the department of Material Science In fulillment of the requirements for the degree of Doktor−Ingenieur (Dr.‐Ing.)

ADissertationapprovedbythedepartmentofMaterialScienceInful illmentoftherequirementsforthedegreeofDoktor−Ingenieur(Dr.‐Ing.)

NanodomainStructureandEnergeticsofCarbonRichSiCNandSiBCNPolymer‐DerivedCeramics

M. Sc. Yan Gao fromXinjiang,ChinaMatrikel‐Nr.1592533

Referee:Prof.Dr.RalfRiedelCo‐referee:Prof.Dr.WolfgangEnsinger

FachbereichMaterial‐undGeowissenschaftenTechnischeUniversitätDarmstadt

Dateofsubmission:09.10.2013Dateoforalexamination:27.11.2013

Darmstadt2014D17

Acknowledgments

The Ph.D. thesis represents the work performed at TU Darmstadt between April 2010 and

June 2013.

I would like to give my sincere thanks to the people who have helped me during this Ph.D.

work:

Many thanks to Prof. Ralf Riedel, who gave me the chance to work in his group and on this

interesting topic. Moreover, warm thanks for his supervision and mentorship during my

research, for his help and support during my entire Ph.D. time, especially when I was sick,

thanks for his great understanding.

Warm thanks to Dr. Gabriela Mera for the suggestions to my work and many interesting

discussions, and help with the Raman measurements.

Lots of thanks to Dipl.-Ing. Hong Nguyen for her assistance with polymer synthesis, and for

giving me help and care in both my work and my life.

Thanks to Prof. Alexandra Navrotsky, Prof. Sabyasachi Sen, Dr. Tien Tran, Dr. Amir

Hossein Tavakoli, Dr. Scarlett Widgeon, Dr. Emil Stoyanov for calorimetry and MAS NMR

support, and for help during my visit at UC Davis.

Thanks to Prof. Hans-Joachim Kleebe and M. Sc. Stefania Hapis, Dipl.-Ing. Mathis M.

Müller for TEM measurements.

Thanks to Dr. Yeping Xu for MAS NMR measurements.

Thanks to Dipl.-Ing. Claudia Fasel for TG/MS measurements and a lot of technical help in

daily lab work.

Thanks to Dr. Koji Morita for TEM investigations.

Thanks to Jean-Christophe Jaud and Dr. Joachim Brötz for the XRD measurements.

Thanks to Dipl.-Ing. Mirko Reinold and Dr. Magdalena Graczyk-Zajac for eletrochemical

study, M. Sc. Mahdi Seifollahi Bazarjani for SAXS measurements, Dipl.-Ing. Jiadong Zang

for impedance spectroscopy measurements, Dr. Tinka Spehr for SAXS measurements, Prof.

Weiyou Yang for TEM investigation, Prof. Sabyasachi Sen for checking SAXS interpretation,

Dr. Magdalena Graczyk-Zajac, Prof. Linan An, Dr.-Ing. Holger Maune and Dr. Ming Li for

the discussion on impedance spectroscopy, Dr. Benjamin Papendorf for helping with

“Zusammenfassung”, M. Sc. Jia Yuan for helping with lab works, Su-Chen Chang for helps

with documentation.

Thanks to my officemates Dr. Magdalena Graczyk-Zajac, Dipl.-Ing. Jan Kaspar, Dipl.-Ing.

Mirko Reinold who are companying me, helping me, comforting me and supporting me.

Thanks to all the DF members who gave me such warm atmosphere and lots of helps. I

cannot accomplish anything without you guys.

Thanks to the Hiwis Shenshen He, Daniel Bick and Senan Jadeed.

I also greatly thank the DFG-NSF project for the financial support.

I give my special thanks to my family: my mother Hong Zhao, my sister Juan Gao, and

my nephew Gaocheng Li. Thank you for being with me all the time no matter what I am and

where I am. I would not be so happy without your support.

Table of contents

Abstract ............................................................................................................................................................ I

Zusammenfassung ......................................................................................................................................... III

1. Introduction and motivation ......................................................................................................................... 1

2. Literature review .......................................................................................................................................... 6

2.1 Synthesis of Si-based preceramic polymers ...................................................................................... 6 2.2 Processing of polymer-derived ceramic bulks ................................................................................... 8 2.3 Energetics of PDCs from calorimetry ................................................................................................ 9 2.4 Nanodomain structure of Si(B)CN ceramics ................................................................................... 11 2.5 Impedance spectroscopy and its use in PDCs .................................................................................. 13 2.6 Electrochemistry of SiCN ceramics as anode material in lithium-ion batteries .............................. 15

3. Experimental .............................................................................................................................................. 17

3.1 Chemicals ........................................................................................................................................ 17 3.2 Synthesis of Si-based preceramic polymers .................................................................................... 17

3.2.1 Synthesis of polysilylcarbodiimides .................................................................................... 19 3.2.2 Synthesis of polyborosilylcarbodiimides ............................................................................ 20 3.2.3 Synthesis of polysilazanes ................................................................................................... 21 3.2.4 Synthesis of polyborosilazanes ........................................................................................... 21 3.2.5 Synthesis of polyphenylsilsesquicarbodiimides .................................................................. 22 3.2.6 Synthesis of polyphenylsilsesquiazane ............................................................................... 23

3.3 Pyrolysis of polymers ...................................................................................................................... 23 3.4 Bulk ceramic processing .................................................................................................................. 24 3.5 Annealing of ceramics ..................................................................................................................... 26 3.6 Characterization techniques ............................................................................................................. 26 3.7 Characterization of ceramics prepared at selected temperatures ..................................................... 29

4. Results and discussion ............................................................................................................................... 33

4.1 Polymer synthesis and processing ................................................................................................... 33 4.1.1 Synthesis of poly(boro)silylcarbodiimides and bulk ceramic processing ........................... 33 4.1.2 Synthesis of poly(boro)silazanes and bulk ceramic processing .......................................... 35 4.1.3 Polysilsesquicarbodiimide and 13C/15N isotope-enriched polysilsesquicarbodiimide ......... 45 4.1.4 Summary ............................................................................................................................. 51

4.2 Effect of processing route ................................................................................................................ 52 4.2.1 Effect of processing route on the nanostructure — Low temperature thermal transformation ...................................................................................................................................................... 52 4.2.2 Effect of processing route on the nanostructure — Thermal transformation up to 2100°C 62 4.2.3 Summary ............................................................................................................................. 68

4.3 Thermodynamic stability ................................................................................................................. 69 4.3.1 Structure and energetics of polysilylcarbodiimide-derived SiCN ceramics ........................ 69 4.3.2 Structure and energetics of poly(boro)silazane-derived Si(B)CN ceramics ........................ 78 4.3.3 Summary ............................................................................................................................. 82

4.4 Solid state structure and microstructure of Si(B)CN ceramics ........................................................ 83 4.4.1 Solid state structure of Si(B)CN ceramics........................................................................... 83 4.4.2 Microstructure of Si(B)CN ceramics .................................................................................. 96 4.4.3 Summary ........................................................................................................................... 103

4.5 Electrical / electrochemical properties ........................................................................................... 105 4.5.1 Complex impedance spectra of selected polymer-derived ceramics ................................. 105 4.5.2 Carbon-rich SiCN ceramics as anode material for lithium-ion batteries ............................ 113 4.5.3 Summary ............................................................................................................................ 115

5. Conclusion ............................................................................................................................................... 117

6. Outlook .................................................................................................................................................... 120

References ................................................................................................................................................... 121

Curriculum Vitae ......................................................................................................................................... 128

Personal status ..................................................................................................................................... 128 Scientific background .......................................................................................................................... 128 Research experiences .......................................................................................................................... 128 Working experience ............................................................................................................................. 129 Activities ............................................................................................................................................. 130 Publications ......................................................................................................................................... 130 Awards & Honors ................................................................................................................................ 131

Eidesstattliche Erklärung ............................................................................................................................. 132

Abstract I

Abstract

This Ph.D. thesis focuses on the synthesis, processing, solid state structure,

nanodomain structure, structural evolution, thermodynamic stability, and functional properties

of carbon rich SiCN and SiBCN ceramics derived from preceramic polymers with tailored

compositions and structures. The main objective of the studies is to better understand the

effects of the composition and structure of the starting precursors, on the behavior of the

resultant ceramics.

First, a set of preceramic polymers with systematically varied compositions and

structures were synthesized. They are linear polysilylcarbodiimides and polysilazanes, and

their boron modified counterparts; branched polysilsesquicarbodiimide and its 13C/15N

isotope-enriched counterpart; and branched polysilsesquiazane. The synthesis of these

polymers was investigated using NMR, FT-IR, Raman and TG/DTG. The results demonstrate

that the obtained precursors exhibit the expected compositions and structures.

Then, the effects of processing route on the thermal stability of PDCs were studied by

comparing bulk ceramics with their powder counterparts. The thermal transformation was

investigated using FT-IR, Raman spectroscopy, XRD, TG/DTG and TEM. The results reveal

that bulk ceramics are more thermally stable than their powder counterparts in terms of

resistance to crystallization and decomposition. The Si(B)CN ceramic powders derived from

poly(boro)phenylvinylsilylcarbodiimide contain α/β-SiC crystallites when heat treated at

1400ºC, while their bulk counterparts prepared at the same temperature remain completely

amorphous. It is also found that bulk ceramics exhibit less weight loss than their powder

analogues at temperatures up to 2100°C, especially in the case of SiCN bulk ceramics. The

higher thermal stability of bulk ceramics as compared with the powder counterparts is likely

due to the powders have greater surface area which enhances the carbothermal reaction and

silicon carbide crystallization. In addition, the boron modification impedes the degradation of

silicon nitride in both bulk and powder samples, similar to previously reported results.

Next, the energetics of PDCs was investigated using high temperature oxidative drop

solution calorimetry on (i) ceramics derived from branched and linear polysilylcarbodiimide;

(ii) ceramics derived from poly(boro)silazanes pyrolyzed at different temperatures. The

results reveal that the ceramics derived from the branched polymer is energetically more

stable than those from the linear polymer. Structural analysis using MAS NMR suggests that

the increased energetic stability of the ceramics derived from the branched polymer is likely

due to the presence of hydrogen at the mixed bonding environments consisting of N, C and Si

atoms. These environments make up the interfacial region between the Si3N4 and “free”

carbon nanodomains. For both the linear and branched polymers, the ceramics derived at 800

ºC are energetically more stable than those derived at 1100 ºC. The study of group (ii) reveals

the effect of the pyrolysis temperature on the structural evolution and energetics of

poly(boro)silazane-derived Si(B)CN ceramics. These ceramics contain mixed SiCxN4-x(x=0-4)

II Abstract

tetrahedra. MAS NMR spectroscopy of the 1100ºC and 1400ºC ceramics reveals that the

structural evolution involves the following processes: (i) the demixing of SiCxN4-x mixed

bonding environments, (ii) the cleavage of mixed bonds at the interdomain regions, and (iii)

the coarsening of domains. Calorimetry results demonstrate that this structural evolution is

favorable in both enthalpy and free energy.

We also investigated the solid state structures and nanodomain structures of the

ceramics derived from the tailored polymers. The MAS NMR results indicate that the SiCN

ceramics derived from polyphenylvinylsilylcarbodiimide and polymethylvinylsilylcarbo-

diimide both contain nanodomains of silicon nitride and “free” carbon. In the ceramics

prepared from the phenyl-containing polymer, the “free” carbon and silicon nitride domains

are basically isolated, while in the ceramics derived from the methyl-containing polymer,

“free” carbon and silicon nitride domains are connected via C-N bonds. The SiBCN ceramics

derived from boron-modified polysilylcarbodiimides have an additional B-containing phase

which is located at the interface between the silicon nitride and the “free” carbon phase. On

the other hand, the SiCN ceramics derived from polysilazanes contain “free” carbon and

mixed bonded SiCxN4-x(x=0-4) nanodomains. The SiCxN4-x (x=0-4) domain consists of a core

of SiN4 tetrahedra, which connect to “free” carbon domain via SiN3C, SiN2C2, SiNC3, and

SiC4 tetrahedra. SiC4 tetrahedra make up the most outer shell of mixed bonded SiCxN4-x(x=0-

4) nanodomains. The SiBCN ceramics derived from boron-modified polysilazanes have an

additional B-containing phase which connects the SiC4 with the “free” carbon domains.

SAXS results indicate that (i) the size of Si-containing nanodomains in SiCN ceramic is larger

than in their SiBCN counterparts; (ii) the ceramics derived from phenyl-containing polymers

exhibit smaller Si-containing nanodomains than those derived from methyl-containing

polymers; and (iii) the size of Si-containing domains increases with pyrolysis temperature in

all ceramics.

Finally, some functional behaviors of the resultant ceramics were investigated. The

electrochemical properties of the ceramics were investigated to explore their applicability as

anode materials for lithium-ion batteries. The results reveal that the SiCN ceramics derived

from both linear and branched polymers are suitable anode materials for lithium-ion batteries.

In particular, the SiCN ceramics derived from linear polyphenylvinylsilazane demonstrate

outstanding performance in terms of cycling stability and capacity at high currents. In

addition, the AC conductivity was characterized using impedance spectroscopy. It is found

that the impedance spectra vary in accordance to the different microstructures of the ceramics,

which, in turn, are related to the chemistry of the precursors. Conduction in SiCN ceramics is

dominated via “free” carbon and SiC phases in series, as analyzed by two semicircles in the

Nyquist plot. Conduction in SiBCN ceramics is dominated by one phase, and is thus

represented by one semicircle in the Nyquist plot. Nonconducting BN forms the interfacial

phase between the “free” carbon and SiC phases, and isolates the discontinuous phase from

the continuous phase. Therefore, conduction in SiBCN ceramics is dominated by the

continuous phase, whether it is “free carbon” or SiC.

Zusammenfassung III

Zusammenfassung

Die vorliegende Dissertation thematisiert die Synthese, Herstellung, Mikro- und

Nanostruktur, strukturelle Entwicklung, thermodynamische Stabilität und funktionellen

Eigenschaften von kohlenstoffreichen SiCN und SiBCN Keramiken, die von präkeramischen

Polymeren mit maßgeschneiderter Zusammensetzung und räumlicher Struktur abgeleitet

worden sind. Das Hauptanliegen der Arbeit ist es, zu ermitteln, inwieweit die

Zusammensetzung und Struktur der polymeren Vorstufen die Eigenschaften der finalen

Keramik beeinflussen.

Zu Beginn wurde eine Reihe von präkeramischen Polymeren mit systematisch

variierten Zusammensetzungen und Strukturen synthetisiert. Dazu gehören lineare

Polysilylcarbodiimide und Polysilazane sowie deren bormodifizierte Äquivalente, verzweigte

Polysilsesquicarbodiimide und deren 13C/15N isotopenangereichertes Pendant als auch

verzweigte Polysilsesquiazane. Die synthetisierten Polymere wurden unter Verwendung von

NMR, FT-IR, Raman und TG/DTG charakterisiert.

Anschließend wurde der Einfluss des Herstellungsverfahrens auf die thermische

Stabilität der polymerabgeleiteten Keramiken untersucht, indem die weitere Aufarbeitung des

Materials entweder als Bulkmaterial oder als Pulver erfolgte. Die thermische Umwandlung

wurde mit Hilfe von FT-IR- und Raman-Spektroskopie, XRD, STA und TEM untersucht. Die

Ergebnisse zeigen, dass die Keramiken als Bulkmaterial thermisch stabiler sind als in Form

von Pulvern in Bezug auf Kristallisationsbeständigkeit und Zersetzung. Die von

Poly(boro)phenylvinylsilylcarbodiimid abgeleiteten keramischen Si(B)CN Pulver enthalten

nach thermischer Auslagerung bei 1400 °C α/β-SiC Kristallite, während das Bulkmaterial

nach Auslagerung unter denselben Bedingungen vollständig amorph bleibt. Zusätzlich wurde

festgestellt, dass die Keramiken als Bulkmaterial nach der Auslagerung bei 2100 °C einen

geringeren Massenverlust als die Pulverkeramiken aufweisen, insbesondere für den Fall einer

SiCN basierten Matrix. Die höhere thermische Stabilität der Bulkkeramiken im Vergleich zu

den Pulverkeramiken ist sehr wahrscheinlich auf die größere Oberfläche der Pulver

zurückzuführen, welche die carbothermische Reaktion und Kristallisation von Siliciumcarbid

begünstigen. Die Modifikation mit Bor verhindert außerdem die Zersetzung des

Siliciumnitrids sowohl im Pulver- als auch im Bulkmaterial, ähnlich wie es bereits in früheren

Publikationen berichtet wurde.

Außerdem erfolgte die thermodynamische Charakterisierung der polymerabgeleiteten

Keramiken mittels oxidativer Hochtemperatur-Tropfenlösungskalorimetrie von (i) Keramiken

abgeleitet von verzweigten und linearen Polysilylcarbodiimiden und (ii) Keramiken abgeleitet

von Poly(boro)silazanen, pyrolysiert bei verschiedenen Temperaturen. Die Ergebnisse zeigen,

dass die von verzweigten Polymeren abgeleiteten Keramiken energetisch stabiler sind als

diejenigen, die auf linearen Polymeren basieren. Strukturanalysen mittels MAS-NMR lassen

vermuten, dass die größere energetische Stabilität der von verzweigten Polymeren

IV Zusammenfassung

abgeleiteten Keramiken sehr wahrscheinlich auf die Präsenz von Wasserstoff und von

gemischten Bindungen – bestehend aus N, C und Si – zurückzuführen ist. Die gemischten

Bindungen bilden die Grenzfläche zwischen Si3N4 und „freien“ Kohlenstoff Nanodomänen.

Unabhängig davon, ob von linearen oder verzweigten Polymeren ausgegangen wird, sind die

Keramiken nach der Pyrolyse bei 800 °C energetisch stabiler als nach 1100 °C. Der Effekt der

Pyrolysetemperatur auf die strukturelle Entwicklung und Energetik der von Poly(boro)-

silazanen abgeleiteten Si(B)CN Keramiken wurde ebenfalls untersucht. Die Matrix dieser

Keramiken besteht aus gemischten SiCxN4-x (x=0-4) Tetraedern. MAS-NMR Spektroskopie

von Keramiken, die bei 1100 °C und 1400 °C pyrolysiert worden sind, zeigt, dass die

strukturelle Entwicklung wie folgt abläuft: (i) eine Phasenseparation gemischter SiCxN4-x

Bindungen, (ii) die Spaltung gemischter Bindungen an den Domänengrenzen und (iii) ein

Domänenwachstum. Kalorimetrische Messungen belegen die Favorisierung dieser

strukturellen Entwicklung in Bezug auf die Enthalpie und freie Energie.

Die Mikro- und Nanodomänenstruktur der polymerabgeleiteten Keramiken wurde

ebenso untersucht. Die MAS-NMR Messungen belegen, dass die von Polyphyenylvinyl-

silylcarbodiimid und Polymethylvinylsilylcarbodiimid abgeleiteten SiCN Keramiken in

beiden Fällen Nanodomänen enthalten, welche aus Siliciumnitrid und „freiem“ Kohlenstoff

bestehen. Werden die Keramiken aus phenylhaltigen Polymeren hergestellt, sind die

Domänen aus „freiem“ Kohlenstoff und Siliciumnitrid grundsätzlich isoliert, wohingegen

diese via C-N Bindungen miteinander verbunden sind, sofern von methylhaltigen Polymeren

ausgegangen wird. Die aus bormodifizierten Polysilylcarbodiimiden abgeleiteten Keramiken

enthalten zusätzlich eine borhaltige Phase, die an der Grenzfläche zwischen Siliciumnitrid

und „freiem“ Kohlenstoff lokalisiert sein kann. Andererseits enthalten die von Polysilazanen

abgeleiteten SiCN Keramiken Nanodomänen von „freiem“ Kohlenstoff und von gemischten

SiCxN4-x (x=0-4) Bindungen. Die SiCxN4-x (x=0-4) Domänen bestehen aus einem Kern von

SiN4 Tetraedern, die mit dem „freien“ Kohlenstoff über SiN3C, SiN2C2, SiNC3, und SiC4

Tetraeder miteinander verknüpft sind. Die äußere Schicht der SiCxN4-x (x=0-4) Nanodomänen

bildet größtenteils SiC4 Tetraeder aus. Die von bormodifizierten Polysilazanen abgeleiteten

SiBCN Keramiken besitzen zusätzlich eine borhaltige Phase, welche die SiC4 und

„freien“ Kohlenstoffdomänen miteinander verbindet. SAXS Messungen belegen, dass (i) die

Größe der Si-haltigen Nanodomänen in SiCN Keramiken größer als in SiBCN Keramiken ist,

(ii) die von phenylhaltigen Polymeren abgeleiteten Keramiken kleinere Si-haltige

Nanodomänen aufweisen als die von methylhaltigen Polymeren abgeleiteten Keramiken und

(iii) in allen Keramiken die Größe der Si-haltigen Domänen mit Erhöhung der

Pyrolysetemperatur zunimmt.

Zuletzt wurde das Funktionsverhalten der resultierenden Keramiken untersucht. Die

elektrochemischen Eigenschaften der Keramiken wurden untersucht, um die Anwendbarkeit

als Anodenmaterial in Lithium-Ionen Batterien zu evaluieren. Die Ergebnisse zeigen, dass die

von linearen und vernetzten Polymeren abgeleiteten SiCN Keramiken als Anodenmaterial für

Lithium-Ionen Batterien geeignet sind. Insbesondere zeigen die von linearen

Zusammenfassung V

Polyphenylvinylsilazanen abgeleiteten SiCN Keramiken eine hervorragende Leistung in

Bezug auf Zyklenbeständigkeit und Kapazität bei hohen Stromstärken. Zusätzlich wurde die

Wechselstrom-Leitfähigkeit mittels Impedanzspektroskopie untersucht. Es wird festgestellt,

dass sich die Impedanzspektren in Übereinstimmung mit den verschiedenen Mikrostrukturen

der Keramiken verändern, die wiederum von der Chemie der polymeren Vorstufen abhängig

sind. Die Leitfähigkeit in SiCN Keramiken wird von „freiem“ Kohlenstoff und SiC Phasen

bestimmt, wie im Nyquist-Diagramm durch das Auftreten zweier Halbkreise belegt wird.

Hingegen wird die Leitfähigkeit in SiBCN Keramiken durch eine einzelne Phase bestimmt, da

die Signalantwort im Nyquist-Diagramm einen einfachen Halbkreis beschreibt. Nichtleitendes

BN bildet die Grenzfläche zwischen dem „freien“ Kohlenstoff und den SiC Phasen und

isoliert die diskontinuierliche Phase von der kontinuierlichen. Daher wird die Leitfähigkeit in

SiBCN Keramiken allein von einer kontinuierlichen Phase bestimmt, unabhängig davon ob es

sich um „freien“ Kohlenstoff oder SiC handelt.

Introduction and motivation 1

1. Introduction and motivation

Studies of polymer-derived ceramics (PDCs) have been undertaken for decades and

have achieved substantial progress in understanding and utilization this unique class of

amorphous ceramics. PDCs possess a set of outstanding properties, e.g. high strength, high

creep resistance, excellent oxidation and corrosion resistance, and outstanding stability

against crystallization and decomposition. Therefore, they are promising materials for

widespread applications, including high strength fibers, coatings, ignition plugs, MEMS

devices, as well as anode materials for lithium-ion batteries [1]. These outstanding properties

are attributed to a unique microstructure of the PDCs, which can only be obtained by

thermolyzing polymeric precursors at moderate temperatures, and cannot be prepared from

ceramic powders using conventional ceramic sintering routes or traditional physical/chemical

deposition processes.

Deeper investigations of PDCs reveal that the X-ray amorphous materials actually

contain various nanodomain structures. The constitution of the nanodomain structures

strongly depends on the material system, the characteristics of the precursors, and the

processing conditions [2–4]. The nanoscale heterogeneity and the characteristics of the

nanodomains have attracted extensive attention because they greatly influence the structure

and properties of the ceramics at a larger/macroscopic scale.

Among all PDCs, SiCN and SiBCN are well known for their outstanding thermal

stability against crystallization and decomposition up to extremely high temperatures. A

variety of polymers, including typical polysilazanes and polysilylcarbodiimides were

developed and used in preparing SiCN ceramics. These polymers were then modified with

boron through, for example, hydroboration reactions to produce SiBCN ceramics. The

microstructure of low-carbon SiCN synthesized from polysilylcarbodiimides is characterized

by the presence of two amorphous phases, namely amorphous Si3N4 and amorphous carbon.

On the other hand, carbon-rich SiCN prepared from polysilylcarbodiimides are composed of

three amorphous phases, Si3N4, SiC and carbon [2,5]. It is found that introducing of excess

amounts of carbon leads to higher thermal stability in SiCN PDCs [5,6]. In SiBCN PDCs,

boron is present as a BCN phase coexisting with the Si3N4, SiC and carbon phases. Studies on

SiBCN ceramics showed that the decomposition temperature of SiBCN is even higher. Some

claimed that the improved stability is due to the increased nitrogen partial pressure in the

vicinity of Si3N4 [7,8], while others believe that it is due to the decrease of carbon activity

because of the dissolution of carbon in BNC layers [7,9]. In both scenarios the equilibrium

temperature of the carbothermal reaction is increased, thus Si3N4 is stable at higher

temperatures.

Recently the use of calorimetry makes it possible to obtain thermochemical data for

PDC materials. It is found that relative to a mixture of crystalline components with the same

composition, the ternary amorphous polymer-derived ceramics have exothermic heats of

2 Introduction and motivation

formation. This phenomenon was found over a large range of compositions in the SiOC,

SiCN, SiCNO and SiBCN systems [10–13], suggesting that the amorphous PDCs are

thermodynamically stable compared to their crystalline counterparts. Such stability is

attributed to the unique structure of the PDCs, particularly the nature of their nanodomains

and interfaces. Studies of certain SiOC PDCs revealed that mixed bonds in the interdomain

regions were the source of the negative enthalpy, perhaps because they experience far less

strain in an amorphous structure than in a crystalline one [14]. Further studies found that

hydrogen release and phase separation could also stabilize the structure [12]. The lack of

mixed bonds reduced the thermodynamic stability in some polysilylcarbodiimide-derived

carbon-rich SiCN ceramics [10]. Moreover, regarding SiBCN which shows extraordinary

resistance to decomposition with the presence of boron, calorimetry data show that

thermodynamic stability diminishes with an increase in boron content [13].

PDCs are fascinating also because the compositions of the ceramics can be tailored by

changing the structure and chemistry of the precursors. Carbon and boron contents have been

the primary focus in studying these ternary and quaternary materials. High amounts of excess

carbon inhibit the crystallization of Si3N4 in SiCN ceramics, leading to greater resistance to

thermolysis than in ceramics containing lower carbon amounts [5]. Boron increases the

stability of SiBCN PDCs against decomposition [13,15,16]. When the two factors exist

concomitantly, a high “free” carbon concentration and boron-modification is assumed to

improve the thermal stability of the SiBCN materials [17–19].

To better understand and use SiCN and SiBCN PDCs, it is important to verify how the

unique nanodomain structures influence the energetics of these PDCs, and to understand the

effect of carbon, boron, etc. on the performance of PDCs.

In this Ph.D. work, two classes of polymers, namely polysilazanes and

polysilylcarbodiimides, as well as their boron-modified counterparts, were designed and

synthesized. In addition, carbon contents were modified by replacing methyl groups attached

to Si atoms with phenyl groups, assuming that the phenyl groups will produce more carbon

than the methyl groups. In this way, the precursors were obtained with systematically varied

molecular architecture, carbon content, and boron modification. The resultant amorphous

SiCN and SiBCN ceramics consisted of Si3N4-SiC-C and Si3N4-SiC-C-BN phase system,

respectively, in terms of stoichiometric components.

Both bulk and powder ceramics were prepared from these precursors. First, bulk

ceramics were prepared from the aforementioned precursors. While PDC route is promising to

produce bulk ceramics by plastic techniques [1,20], the main challenge is the evolution of gas

during pyrolysis and the accompanied shrinkage, deformation, pore formation and cracking.

There are various methods for forming bulk ceramics [21–32], and warm-pressing is one of

the most commonly used and effective routes. For the successful usage of this technique,

processing parameters, e.g. pressing pressure, heating temperature, and holding time must be

optimized in accordance with the features of each precursor. In a typical warm-pressing,


concomitant polymer crosslinking is required for obtaining the greenbody shape. Extra pre-

crosslinking is needed for liquid (honey-like) polymers i.e. polysilylcarbodiimides and

polysilazanes. The parameters for this essential step must be optimized in order to not only

produce solidified pre-crosslinked polymer for the subsequent warm-pressing, but also to

preserve certain reactivity in the warm-pressing procedure. For the polyborosilazanes

synthesized in this work, bulk samples cannot be prepared directly by warm-presssing due to

their thermoplastic behavior. Therefore, the self-filler route was introduced. The preparation

of the self-filler and the ratio between the original polymer and the fillers are also required to

be optimized.

PDC powders were also prepared in this work. Obviously, bulk ceramics have much

lower surface area than their powder counterparts. Previous studies on the effect of particle

size found that for SiBCN ceramics, increasing surface areas enhanced the carbothermal

reaction, thus reduced the crystallization temperature of the materials [33,34]. It was found

that poly(boro)phenylvinylsilylcarbodiimide-derived carbon-rich SiCN and SiBCN bulk

ceramics exhibited considerably different microsctures as compared to the powder samples.

The final microstructure, crystallinity, and resistance to thermal decomposition are strongly

influenced by the precursor chemistry, the processing route, the thermolysis temperature, the

presence of boron, and the “free” carbon content of these materials. Therefore, the properties

of PDCs and their potential applications depend, to a great extent, on the chemistry of the

precursors and their processing history.

Thermodynamic stability is the physiochemical reason for the formation of certain

structures. Two groups of samples were selected for calorimetry measurement, namely i)

branched polysilsesquicarbodiimide- and linear polysilylcarbodiimide-derived SiCN ceramics

pyrolyzed at 800ºC and 1100ºC, and ii) polysilazane-derived SiCN and polyborosilazane-

derived SiBCN ceramics pyrolyzed at 1100ºC and 1400ºC. The first group was studied to

clarify the effect of polymer structure, i.e. branched versus linear, on the energetics of the

derived ceramics. The second group was selected to understand the energetic of the

microstructural evolution as a response to heat treatment, as well as the role of boron in this

respect. With the help of MAS NMR spectrocopy, the structures of these ceramic samples

were comprehensively evaluated. The compositions of the nanodomains as well as the

interdomain regions were probed in order to better interpret differences in the energetic

stability of the materials.

Since different precursors result in very different nanodomain structures, the

overarching aim of this work is to develop structural models for ceramics derived from

tailored polymers. Therefore, characterizing the structures becomes very important. In this

work, many powerful techniques were utilized to characterize the structures of the resultant

ceramics at different length scales, including MAS NMR, XRD, Raman, TEM, and SAXS.

The integration of all of these methods enables a deep understanding of the nanodomain

structure in terms of domain type, domain size, chemical bonding at domain interfaces, etc.

4 Introduction and motivation

The ceramic structures show strong dependance on their precursors. Considering their varied

molecular architectures, carbon content, and boron presence, different polymers result in

ceramics with unique features. This phenomenon can be explained by the nature of the

starting polymers, as well as the polymer to ceramic transformation.

With a comprehensive understanding of the nanodomain structures in these ceramics,

the electrical properties were finally characterized by impedance spectroscopy (IS). IS is a

very sensitive electrochemical technique which allows the simple treatment of carrier

transport processes and reactions in complex situations, such as in intrinsically disordered

material components [35]. The greatest benefit of using IS to study the conduction behavior

of PDCs is the possibility to distinguish the effects of different phases. For this reason, it is

also a commonly used technique for measuring the electrical properties of multiphase

ceramics. There have been a few reports using impedance spectroscopy [36,37] to study the

AC conductivity (σac) and the mechanisms of relaxation in PDCs. However, these studies did

not emphasize the importance of the nanodomain structure. To date, no report has yet

distinguished the σac contribution from individual types of nanodomains in the structure of

PDCs. Based on the results from IS, the contribution of the different phases on the

electrochemical behavior of PDCs is interpreted and discussed.

PDCs are also promising as anode materials for lithium-ion batteries. Their capacities

exceed that of graphite, and they also exhibit stable cycling behavior. Ahn et al. [38] claimed

that it is necessary for the nitrogen to oxygen (N/O) ratio to be below 1 to achieve high

capacities of ca. 600 mA h g-1. In this work, four SiC(O)N ceramics with a N/O ratio of 6 to

9.5 show outstanding electrochemical properties. These four ceramics were derived from

polymers with different molecular structures and degrees of branching, namely linear and

branched polysilylcarbodiimides and polysilazanes. The focus of this study is the relationship

between the molecular structure of the precursor on the microstructure of the derived ceramics,

and thus on the electrochemical properties of the materials.

In summary, this Ph.D. work demonstrates that the structures of polymer-derived

ceramics can be widely varied by tailoring the precursors in terms of molecular architecture,

boron modification, and carbon contents. It also demonstrates that the behavior of the

obtained materials can be strongly affected by processing route (i.e., for powder vs. bulk). It

was found that polymer-derived ceramics are thermodynamically more stable than their

crystalline constituting phases. The origin of the thermal stability was attributed to the

nanodomain structure of the amorphous PDCs. Models of the nanodomain structures were

constructed for all systems, taking into account nanodomain types, nanodomain size and

bonding environment at the interdomain regions. The effect of this unique nanodomain

structure on the electrical properties of PDCs was studied, and impedance spectroscopy was

employed to distinguish the contribution of different phases to the electrical behavior. The

electrochemical study demonstrates that certain SiCN ceramics with a N/O ratio of about 6-


9.5 show outstanding behavior in light of their potential application as lithium-ion batteries

anodes.

6 Literature review

2. Literature review

Polymer-derived ternary SiCN and quaternary SiBCN ceramics are refractory

materials exhibiting a great resistance to crystallization and decomposition at high

temperatures [1,39,40]. Unlike conventional polycrystalline ceramics made by sintering of

corresponding powders, polymer-derived ceramics possess a unique structure that contains a

Si-based amorphous phase, and in some cases, a highly-disordered “free” carbon phase. While

PDCs are X-ray amorphous, they are nanoscopically heterogeneous, containing intrinsically

complex nanodomains, which can be varied in a large range by tailoring the precursor

chemistry and processing conditions. The properties of PDCs also strongly depend on

precursor chemistry and processing conditions [1,20,40,41]. For example, compared to low

carbon SiCN systems, previous studies showed that carbon-rich SiCN ceramics have much

higher stability against crystallization and remain amorphous up to higher

temperatures [2,5,10]. This phenomenon was attributed to the finely dispersed “free” carbon

phase which acts as a diffusion barrier in the structure. The presence of “free” carbon and

boron dramatically influences the properties of the ceramics with respect to their

microstructure. Carbon-rich polymer-derived SiBCNs are defined as complex nanostructured

systems which have remarkably high thermal, chemical, and mechanical stability (i.e., creep

resistance) even at temperatures up to 2000-2200 °C in inert atmospheres [17,42].

The best-known precursors for silicon carbonitride (SiCN) ceramics are polysilazanes

-[R1R2Si-NR3]n- and polysilylcarbodiimides -[R1R2Si-N=C=N]n-, where R1, R2 and R3 are

organic groups such as methyl, phenyl, etc. As demonstrated by solid-state NMR

studies [1,20,40], ceramics derived from polysilazanes contain mixed bonding environments

about the Si atoms forming SiCxN4-x (x = 0-4) tetrahedra. Meanwhile, the ceramics derived

from polysilylcarbodiimides contain two amorphous phases, namely, amorphous Si3N4 (a-

Si3N4) and amorphous carbon (a-C). The pyrolysis of highly carbon-rich

polysilylcarbodiimides produces a microstructure composed of three amorphous phases— a-

Si3N4, a-SiC and a-C [2,5].

Polymer chemistry and processing routes play important roles on the nanodomain

structure, thereby, the performance of SiCN and SiBCN ceramics. Here, SiCN and SiBCN

PDCs were studied based on previous findings. This chapter addresses previous

accomplishments in polymer synthesis, bulk processing, energetic stability, nanodomain

structure, and the electrochemical study of SiCN and SiBCN PDCs. This review offers a

general understanding of the field, and highlights the contributions of this work to the current

understanding of PDCs.

2.1 Synthesis of Si-based preceramic polymers

Polysilylcarbodiimides are an important class of precursors for SiCN ceramics. The

popularity of these precursors grew after many previous studies revealed that

Literature review 7

polysilylcarbodiimides-derived ceramics were more thermally stable than analogous ceramics

derived from polysilazanes [6,39,43–48]. Bis(trimethylsilyl)carbodiimide and cyanamide are

typical starting materials for the synthesis these polymers, as shown in Scheme1 [1]. They are

reacted with organo element halides or pure element halides to form polysilylcarbodiimides.

Scheme1 carbodiimidolysis and cyanamidolysis of chlorosilanes with bis(trimethylsilyl)carbodiimide and

cyanamide [40]

It was found that bis(trimethylsilyl)carbodiimide was an efficient starting material for

synthesizing other element carbodiimides [49–53]. These days, the synthesis of

polysilylcarbodiimides is commonly accomplished by the reaction of chlorosilanes and

bis(trimethylsilyl)carbodiimide with pyridine as a catalyst, following the work of Riedel et

al. [20]. The by-product, trimethylchlorosilane, was removed by vacuum drying. Depending

on the chlorosilanes used, cyclic, linear, or highly-branched poly(silsesquicarbodiimides) can

be obtained [43,44,46,54,55]. Recently, Mera et al. [5] reported a group of linear carbon-rich

polysilylcarbodiimides which contain phenyl functional organic groups, and another

alternative group in the form of hydrogen, vinyl, or methyl. It was found that highly carbon-

rich polysilylcarbodiimide-derived ceramics exhibited even greater thermal stability against

crystallization.

The synthesis of polysilazane precursors could be achieved from chlorosilanes, by

means of ammololysis reactions with ammonia, or aminolysis with different amines. Scheme2

shows the ammonolysis and aminolysis reactions of dichlorosilanes resulting in the synthesis

of polysilazanes. Numerous studies on polysilazanes as ceramic precursor have been carried

out since the 1960s [56–65]. The main disadvantage of these routes is the difficulty of

separating the by-products, NH4Cl or H3NRCl. In the routes shown in Scheme2, chlorosilane

could also be replaced by various silanes which have Si-H groups, and these Si-H groups

could react with N-H groups [58,66–68]. This alternative reaction avoids producing the

chlorine-containing salt by-product. However, silanes cost much more than chlorosilanes, and

silanes are more dangerous to handle. Therefore, silanes are not preferred for obtaining

polysilazanes.

Scheme2 Synthesis routes to polysilazanes starting from chlorosilanes [40]

8 Literature review

In contrast to the typical ammonolysis or aminolysis way for synthesizing polysilazane

precursors, a reflux method was also reported for the synthesis of polysilazanes using

hexamethyldisilazane and chlorosilane as starters, either without catalyst [69,70] or with

AlCl3 [70] as the catalyst.

In this work, a new method using hexamethyldisilazane and chlorosilane as starting

materials and pyridine as catalyst was performed, and polysilazanes with different substitutes

were acquired. This new route has not yet been reported.

2.2 Processing of polymer-derived ceramic bulks

The polymer-derived ceramic route can be used to produce complex-shaped bulk

ceramics using the advantages of plastic technologies [1,20,40]. However, precursor

pyrolysis is always accompanied by the evolution of gaseous by-products, pore formation,

and significant shrinkage which can introduce stresses and deformation. Many efforts were

made to obtain dense bulk ceramics, including pressing, extrusion, and injection molding.

Among these techniques, pressing was most performed and studied. Pressing techniques

includes cold isostatic pressing, cold uniaxial pressing, warm uniaxial pressing, hot isostatic

pressing, etc. These works have been previously summarized elsewhere [40].

Uniaxial warm press is an effective technique for bulk shaping, and is correlated to the

polymer crosslinking process. During warm-pressing and the accompanied crosslinking

process, the shaping of greenbody compacts is achieved by producing duroplastic materials.

The other benefit of crosslinking is that it helps to avoid the loss of oligomers during

pyrolysis, and increases the ceramic yield. Warm presssing is proved to be practical in a

variety of precursor systems [21,23–32,71] when the parameters applied, including

temperature, pressure, presssing time, etc. were optimized. Optimal processing parameters of

course depend on the type of precursor. Pressing temperatures usually vary from 120°C to

320°C, and applied pressure range from 20 to 710 MPa. For liquid polymers, a pre-

crosslinking step is essential in order to obtain dense materials. Pre-crosslinking is considered

as a partial crosslinking, which is to say, the polymers are solidified, but reactive groups

remain for further crosslinking during shaping.

Fillers are often introduced for monolith forming in order to minimize shrinkage of the

whole composite. Passive fillers include metal oxides, carbides or nitrides that occupy space,

but do not react during pyrolysis, and reduce the volume fraction of polymer. Active fillers

are usually metals or intermetallics, e.g. Al, Ti, Zr, B, Si, which react with decomposition

products or in the pyrolysis atmosphere to form a new phase that expends in volume [72–75].

Self-fillers are thermolyzed from the same precursor as the matrix material, and are heat

treated to an intermediate state between the precursor polymer and the final ceramic. Self-

fillers play the same role as normal fillers, but still generate a “pure” system. Fillers can

constitute the volume fraction majority of the final ceramic part, and assist in achieving higher

densities.

Literature review 9

A flow chart of the pressing technique for polymer-derived ceramic processing is

given in Figure 1. The alternative to using fillers is also shown [40].

Figure 1 Pressing technique used in shaping polymer-derived ceramics, “P” marked when pressure is applied. Reprinted with permission of DEStech Publications, Inc., from Polymer Derived Ceramics, Eds. P. Colombo et al. Copyright 2010 [40].

2.3 Energetics of PDCs from calorimetry

The most remarkable property of Si(B)CN PDCs is their extraordinary thermal

stability against crystallization and decomposition. Observations of some SiBCN PDCs reveal

that no significant mass loss occurs at 1600°C, and no tremendous decomposition occurs up

to 2000°C when the Si3N4 phase, which decomposes at 1841°C, is present [42,76–80]. Both

SiCN and SiBCN PDCs have crystallization temperatures above 1400°C. This is a substantial

improvement considering that the crystallization temperature of amorphous SiC is about

1000°C, and that of Si3N4 is about 1200°C [81,82].

Due to the outstanding thermal stability of PDCs, numerous studies have been

conducted to sought out an explanation for why these ternary and quaternary systems possess

such strong resistance to thermal treatment [8,42,80,83–91]. Calorimetry may paint the real

thermodynamic landscape of these materials. High temperature oxidative drop solution

calorimetry is well established [92,93], and has been used to study polysiloxanes-derived

silicon oxycarbide ceramics (SiOC PDCs) [12,14], polysilylcarbodiimides-derived SiCN

ceramics [10,11], and polyborosilazane-derived SiBCN ceramics [13]. Using this method, a

pellet of pressed powder a few mg in mass is dropped from room temperature into a molten

oxide solvent at high temperature in a Tian-Calvet twin microcalorimeter under oxidizing

10 Literature review

conditions (Figure 2). When the pellet is dissolved in sodium molybdate (3Na2O·4MoO3)

solvent, the final state of the sample is SiO2 (cristobalite) and evolved gases (CO2 and

N2) [11,94]. In order to accelerate sample dissolution and maintain an oxidizing environment,

oxygen gas is bubbled through the melt. The evolved CO2 and N2 gases produced by the

dissolution of the sample in the oxide melt are removed by continuously flushing oxygen gas

in the headspace above the solvent. The calorimeter is calibrated using the heat content of Pt

wire [92,93,95].

Figure 2 schematic of high temperature drop solution calorimetry

The results of high-temperature drop solution calorimetry showed that SiOC ceramics

are thermodynamically stable with respect to the crystalline components, namely cristobalite,

graphite and silicon carbide of the same composition. The interfacial region between

nanodomains was claimed to be the source of the negative enthalpy. Mixed bonds in the

interfacial region are far less strained in an amorphous structure than in a crystalline one [14].

Further study of SiOC PDCs produced at different pyrolysis temperatures revealed that, at

lower temperatures, hydrogen releases some strain in the structure, leading to a lower energy

state. At much higher temperatures, the carbon nanodomains gain stability due to phase

separation. The carbon domains grow from single graphene sheets into stacks of graphite-like

layers [12]. Similar interpretations have been made in previous studies of SiOC PDCs based

on ab initio calculations that indicate that the presence of mixed Si−C−O bonds energetically

stabilized the structures of these materials [96].

Studies of carbon-rich polysilylcarbodiimides-derived SiCN PDCs show that these

SiCN PDCs are less stable than the SiOC ceramics studied previously. They have slightly

positive heats of formation from the binary crystalline components Si3N4, SiC, and C [10].

Microstructural studies revealed that significant mixed bonds did not exist in these ceramics,

Literature review 11

but separate Si3N4, SiC and amorphous carbon nanodomains did [2,5]. Tavakoli et al. studied

a series of amorphous Si(B)CN PDCs and demonstrated that thermodynamic stability

diminished with an increase in boron content, meaning the hindering effect of boron on Si3N4

crystallization is kinetic [13,97]. In this work, we study the effect of (i) branched and linear

polymer, (ii) different pyrolysis temperatures, on the the structure and energetics of selected

ceramics.

2.4 Nanodomain structure of Si(B)CN ceramics

Compared to the conventionally processed SiC and Si3N4 ceramics, PDCs can be

synthesized with highly tunable compositions to produce unique amorphous nanodomain

structures. While distinct compositions show very different properties, similar compositions

derived from different precursors or pyrolyzed at different temperatures may also have

dramatically different properties due to variations in the size and nature of their nanodomains

and interdomain regions [6,20,98].

The constitution of nanodomains in these amorphous PDCs depends on the materials

system. Carbon-rich Si-C-N PDCs with compositions located within the limit of the SiC-

Si3N4-C three phase region of the ternary Si-C-N system generally consist of sp2 C

nanodomains and silicon carbonitride domains, the latter composed of SiCxN4-x (x = 0– 4)

tetrahedral units [99–103]. Si-B-C-N PDCs with compositions located in the SiC-Si3N4-C-

BN four phase region of the quaternary Si-B-C-N system generally consist of sp2 C

nanodomains, domains of SiCxN4-x (x = 0–4) tetrahedra, and nanosized features containing B,

N and C [104–106]. During pyrolysis, the covalent bonds in PDCs can rearrange and

redistribute. Accordingly, two phenomena are usually observed, namely coarsening and

demixing. Solid-state NMR investigations of Si-C-N and Si-B-C-N PDCs pyrolyzed at

temperatures above 1000 °C show that demixing of tetrahedral environments in SiCxN4-x

matrices occurs [99,104,106].

In small angle X-ray scattering (SAXS), X-rays are used to investigate the structural

properties of solids, liquids or gels. Photons interact with electrons, and provide information

about the fluctuations of electronic densities in heterogeneous matter. A typical experimental

set-up is shown in Figure 3. The scattered intensity is collected as a function of the so-called

scattering angle 2. Elastic interactions are characterised by zero energy transfers, such that

the final wave vector kf is equal in modulus to ki. The relevant parameter to analyse the

interaction is the momentum transfer or scattering vector q=ki-kf, defined by q=(4/)sin.

The standard unit for q is Å-1.


Figure 3 Schematic view of a typical scattering experiment

The number of photons scattered by a sample is proportional to its total volume V and

to its electronic contrast. The higher the contrast between particles and solvent, the more

intense the scattered signal. Figure 4 shows a binary sample and "q-window" corresponding to

a measurement at a given q0. The contrast is equal to zero in Cases 3 and 4, and non-zero in

Cases 1 and 2.

Figure 4 An example for binary system

Intuitively, a measurement made at a given q0 allows one to investigate the density

fluctuations in a sample based on a distance scale D0=2/q0. It is equivalent to observing the

system through a 2p/q0 diameter "window" in real space, as shown in Figure 4, with the circle

being the observation window. A scattering signal is observed if the contrast Δρ inside the

circle is non-zero. To study objects much smaller or much larger than D0=2/q0, another

"window" must be selected. The q-range of small angle experiments is usually divided into

three main domains, as shown in Figure 5.

Figure 5 q-range of small angle experiments

At a high q domain, the window is very small, and there exists contrast only at

the interface between two media. This domain, called the Porod's region, gives information

about the interfaces present in the sample. In the intermediary zone, the window is of the


order of the "elementary bricks" in the systems. The form factor P(q) can be measured, as it

relates to the size, shape and internal structure of one particle. At a low q domain, the

observation window is very large and the structural order can be obtained through the so-

called structure factor S(q), which allows one to calculate the interactions within the system.

Guinier plots of small angle X-ray scattering (SAXS) data show that the domain size

in Si-(B-)-C-N PDCs increases in the range of ~ 0.5 nm to 4 nm with increasing pyrolysis

temperature and time [107–110].

2.5 Impedance spectroscopy and its use in PDCs

Polymer-derived ceramics thermolyzed possess amorphous structures and outstanding

properties, such as thermal stability against decomposition and crystallization, as well as

interesting optical and electrical properties. Previous studies of electrical properties were

carried out basically focusing on the DC conductivity (σdc) of the materials. Studies of some

polysilazane-derived ceramics [36] found that, for PDCs pyrolyzed/annealed from 1000 ºC to

1400 ºC, σdc increased by as much as 3 orders of magnitude with increasing temperature. This

may due to the loss of residual hydrogen accompanied by an increase of the sp2/sp3-carbon

ratio. At higher temperatures (T > 1400ºC), the pronounced increase in conductivity was

attributed to both the formation of nano-crystalline SiC, and to a reduction in the nitrogen

content of the amorphous matrix. At even higher temperature T > 1600ºC, the SiC particles

formed percolation paths throughout the sample. At annealing temperature above 1300 ºC, the

formation of crystalline SiC and Si3N4, and hence the increase in conductivity, cannot be

considered an intrinsic property of the SiCN ceramics. Other studies found that at pyrolysis

temperature above 1000 ºC, graphitic-like carbon was shown to be the main conductive

phase [101,103].

Despite identifying the main conductive phases in the SiCN system, all experimental

data showed an increase of σdc with increasing annealing temperature. This may be correlated

to the growth of dehydrogenated aromatic carbon and its local arrangement as cage-like

structures around SiC-crystallites [37], as well as with an enhanced sp2/sp3-carbon ratio [36].

A number of experimental investigations [3] and theoretic calculations [111] showed the

existence of cross-linked graphitic carbon networks in the SiCN system. If the content of the

graphitic carbon phase reached the percolation threshold, the electrical properties of the

material are dominated by this phase. This phase may be assumed to be in the form of carbon

clusters, or a network of graphitic-like lamellae a few atom layers in size. In the case of

carbon clusters, about 16 vol% was the threshold for percolation [37], and in the case of

graphitic-like carbon layers, high conductivity could be achieved by a very small volume

fraction (less than 1 vol%) [112]. Samples with high carbon contents obtained at high

temperatures showed Arrhenius dependence [113–115], behaved as if containing graphitic-

like amorphous carbon [116].


Impedance spectroscopy (IS) allows one to simply treat carrier transport processes and

reactions in complex situations, such as in intrinsically disordered material components [35].

Some investigations using impedance spectroscopy [36,37] studied the σac of PDCs, and the

mechanism of relaxation. Impedance spectroscopy is very sensitive to every phase in the

structure, and thus gives the σac of each individual type of nanodomain.

Impedance spectroscopy studies the properties that influence the conductivity of

materials systems intrinsically and/or under external stimulus. The parameters derived from

IS spectrum fall generally into two categories: (a) those pertinent only to the material itself,

such as conductivity, dielectric constant, mobility of charge carriers, equilibrium

concentrations of the charged species, and bulk generation–recombination rates; and (b) those

pertinent to an electrode–material interface, such as absorption-reaction rate constants,

capacitance of the interface region, and diffusion coefficient of neutral species in the electrode

itself [35].

Generally, the experimental impedance data may be approximately simulated by an equivalent circuit made up of resistors, capacitors, inductors and other circuit elements. A resistance (R) represents a conductive path for the bulk conductivity, a dissipative component of the dielectric response, or a chemical step associated with an electrode reaction. In the same way, capacitors (C) and inductors (L) model space charge polarization and electrocrystallization processes at an electrode. The complex impedance of a series circuit RC (Formula 1, Figure 6(a)) and of a parallel circuit (RC) (Formula 2, Figure 6(b)) are given by:

/

Formula 1

Formula 2

(a) RC (b) (RC)

Figure 6: Circuit diagram of serial (a) and parallel (b) R and C

S. Wansom et al. [117] developed a “universal equivalent circuit model” to describe the impedance behavior of composites with conduction particles as shown in Figure 7. Each box represents a parallel resistor and capacitor. Basically, below the percolation point of the particles, the upper two paths agree between the simulated and measured spectra. The uppermost path is the matrix path, where the first element represents the unreinforced matrix. The second element only exists with the addition of insulating particles in the matrix. The middle path accounts for particle path, with the first element representing the electrical properties of the particle. The second element corresponds to the inter-particle current flow at high frequencies, which is usually responsible for the low resistance at high-frequency cusp.


The third element of the middle path and the rightmost element in Figure 7 account for the double layer, respectively, of the particle and external electrode surfaces. The bottom-most path is required to describe the composite beyond the percolation threshold of the particles.

Figure 7: A universal equivalent circuit model for composite with particles/fibers distribution. Reprinted from

Cement and Concrete Composites, Vol 28, S. Wansom, N. J. Kidner, L. Y. Woo, and T. O. Mason, AC-

impedance response of multi-walled carbon nanotube/cement composites, 509-519, Copyright (2006), with the

permission from Elsevier [117].

2.6 Electrochemistry of SiCN ceramics as anode material in lithium-ion batteries

Recently polymer-derived ceramics were also applied as anode materials in lithium-

ion batteries. PDCs exhibit capacities exceeding that of graphite and furthermore show stable

cycling behavior, even at elevated charge/discharge rates. There exists a clash of opinions

with respect to the lithium storage sites in PDCs. Ahn et al. claim that the mixed bonding

configuration (tetrahedrally coordinated silicon from SiC4 via SiC3O, SiC2O2 and SiCO3 to

SiO4) of SiOC ceramics acts as a major lithiation site [38], while Fukui et al. found that the

“free” carbon phase within these materials to provide the major hosting sites for Li ions [118].

So far, considering pure PDCs, most existing publications related to the electrochemical

performance of PDCs have been focused on SiOC ceramics [38,118–121]. Even though Dahn

et al. have already patented the use of silazane-derived SiCN ceramics showing reversible

discharge capacities up to 560 mA h g-1 in 1997 [122], little research has been done on the

application of these materials to lithium-ion batteries since that time. Pure PDC-based SiCN

materials derived from polysilylethylenediamine have been investigated by Su et al. [123]

and Feng [124]. The work of Su et al. showed a first discharge cycle capacity of 456 mA h g-

1, but the material suffered strong fading with cycling. The problem of capacity fading was

solved by Feng, who achieved capacities higher than 300 mA h g-1 after 30 cycles for a

current rate of 160 mA g-1, after an additional heat treatment was applied to the polymer-

derived SiCN material. Promising electrochemical results (capacity and stability) for SiCN

derived from high-carbon containing polysilylcarbodiimides have been reported by Kaspar et

al. [125] and Graczyk-Zajac et al. [126]. Recently, it has been found that composite anode

materials composed of SiCN and graphite [127] and SiCN and silicon [128] demonstrate

better electrochemical properties than that of pure graphite and silicon, respectively. In


parallel, Ahn et al. [129] suggested that the presence of nitrogen in SiOC ceramics strongly

diminishes capacity reversibility. In the above report, it is claimed that high capacities of 600

mA h g-1 can only be found in materials with a nitrogen to oxygen (N/O) ratio below 1. The

results are discussed in relation to the degree of bond covalency, with Si-N bonds being more

covalent than Si-O bonds. It is suggested that lithium is sequestered in Si-O-C mixed bond

environments. Accordingly, the replacement of oxygen by nitrogen within the amorphous

network leads to a loss in the ability of these mixed bonds to reversibly bind lithium. In the

present, work we demonstrate that SiC(O)N materials with a N/O ratio of about 6-9.5 show

outstanding electrochemical properties in view of their potential application as lithium-ion

battery anodes. Furthermore, the molecular structure of the polymer precursors and the

nanostructure of the final ceramics play important roles in the electrochemical behavior of

these polymer-derived ceramics.

Experimental 17

3. Experimental

3.1 Chemicals

The chemicals used here are phenylvinyldichlorosilane, methylvinyldichlorosilane,

phenyltrichlorosilane, hexamethyldisilazane, borane dimethylsulphide, dicyandiamide,

cyanamide, 13C/15N-isotope cyanamide and pyridine. They were purchased from Sigma-

Aldrich Chemie GmbH, Germany. All chemicals were used as-received without further

purification.

3.2 Synthesis of Si-based preceramic polymers

All polymers in this study were house-synthesized at TU Darmstadt. The structures of

the different polymers are shown in Figure 8. Synthesis was performed using the Schlenk

technique in protective argon. All glassware involved was dried under vacuum heating, and

then filled with argon. The precursors and derived ceramics were handled, stored, and

prepared for characterization in a glovebox.

18 Experimental

Figure 8 Description of the preceramic polymers synthesized in this work

Experimental 19

3.2.1 Synthesis of polysilylcarbodiimides

The synthesis of polysilylcarbodiimide was completed using self-made

bis(trimethylsilyl)carbodiimide and pyridine as the catalyst. Dichlorosilane with the vinyl

substituent was used for the synthesis. To vary the carbon content of the precursors, a phenyl

substituent (-C6H5) was used for higher carbon content precursors and a methyl group

substituent (-CH3) was used for lower carbon content precursors. The production of

polysilylcarbodiimide using bis(trimethylsilyl)carbodiimide and chlorosilanes has the

advantage that the byproduct Me3SiCl is easy to remove under vacuum. The syntheses are

described in detail as following.

Bis(trimethylsilyl)carbodiimide was synthesized according to the literature [130], as

shown in Scheme 3. First, dicyandiamide (42 g) was mixed with an excess of

hexamethyldisilazane (177g), and the catalyst ammonium sulfate (0.2 g). The mixture was

constantly stirred at 170°C under refluxing. The created by-product was NH3, which is

released as gas. The mixture was then distilled by a Vigreux column and the phase

bis(trimethylsilyl)carbodiimide was separated from the mixture.

Scheme 3 Reaction of Bis(trimethylsilyl)carbodiimide

Polyphenylvinylsilylcarbodiimide was synthesized using phenylvinyldichlorosilane

and bis(trimethylsilyl)carbodiimide with pyridine as a catalyst (Scheme 4). The synthesis was

performed in inert argon atmosphere. First, bis (trimethylsilyl)carbodiimide (0.24 mol) and

pyridine (0.12 mol) were added into a flask under constant stirring. Then,

dichlorovinylphenylsilane (0.24 mol) was added to the mixture. The mixture was first

refluxed at 66 °C for 7 hours in order to prevent the evaporation of phenylvinyldichlorosilane,

which has a boiling point of 84-87°C. A secondary refluxing was carried out at 120 °C for

additional 11 hours. NMR investigations were carried out to check the reactions. The by-

product, Me3SiCl, was removed under vacuum while heating the mixture at 120 °C.

Polyphenylvinylsilylcarbodiimide was obtained after the separation of the by-product. The

yield of dried polymer was about 85%.

Scheme 4 Synthesis of polyphenylvinylsilylcarbodiimide

20 Experimental

The reaction of polymethylvinylsilylcarbodiimide was performed under the same

conditions as polyphenylvinylsilylcarbodiimide. The reaction is shown in Scheme 5. The

removal of the by-product was carried out at room temperature. NMR measurements were

performed to check the formation of the polymer and the complete removal of the by-product.

The yield of polymer was about 70%.

Scheme 5 Synthesis of Polymethylvinylsilylcarbodiimide

3.2.2 Synthesis of polyborosilylcarbodiimides

Polyborosilylcarbodiimides were synthesized using the polysilylcarbodiimides

prepared above and boranedimethylsulfide (BH3·SMe2). Toluene was used as the solvent and

was removed after reaction.

Polyphenylvinylsilylcarbodiimide (0.24 mol) and toluene (100 ml) were mixed in a

flask in a cold bath consisting of dry ice and isopropanol. After one hour of cooling in the

cold environment under stirring, 0.082 mol boranedimethylsulfide (BH3·SMe2) dissolved in

50 ml toluene was dropped into the mixture from a dropping funnel. When the dropping

finished, the cold bath was removed and the mixture was slowly warmed up to room

temperature. The mixture was then solidified. This hydroboration reaction is shown in

Scheme 6. To remove the toluene, the polymer was dried under vacuum for 3 hours at room

temperature, followed by 9 hours at 66 °C and 7 hours at 120 °C. The reaction had a yield of

100% and the obtained polymer was free from residual toluene.

Scheme 6 Synthesis of polyborophenylvinylsilylcarbodiimide

The synthesis of polyboromethylvinylsilylcarbodiimide was performed using the same

conditions as the polyborophenylvinylsilylcarbodiimide. Polymethylvinylsilylcarbodiimide

was reacted with BH3·SMe2. The synthesis route is shown in Scheme 7.

Experimental 21

Scheme 7 Synthesis of polyboromethylvinylsilylcarbodiimide

3.2.3 Synthesis of polysilazanes

The synthesis of polyphenylvinylsilazane was performed by reacting 1 mol

hexamethyldisilazane (HMDS) and 1 mol phenylvinyldichlorosilane with pyridine (0.5 mol)

as the catalyst, as shown in Scheme 8. The mixture was refluxed at 60°C for 21 hours

followed by a distillation under 120 °C to remove the by-product, trimethylchlorosilane. In

order to further remove trimethylchlorosilane, the product was dried under vacuum at room

temperature for another 5 hours. NMR confirmed that the polymer formed and was free from

the by-product.

HNSi

CH3

CH3

H3C Si NHn

Si

CH3

CH3

H3C ClSi

CH3

CH3

CH3Si ClCl 2nnnPyridine

Scheme 8 Synthesis of polyphenylvinylsilazane

Similarly, the synthesis of polymethylvinylsilazane was carried out by using

methylvinyldichlorosilane and hexamethyldisilazane (HMDS) as the starting materials, and

pyridine as the catalyst, as shown in Scheme 9. The mixture was refluxed at 60°C for 7 h,

120°C for 7h and 180°C for 7h. To remove trimethylchlorosilane, the product was dried under

vacuum at 60°C for 5 hours.

Scheme 9 Synthesis of polymethylvinylsilazane

3.2.4 Synthesis of polyborosilazanes

Polyborosilazanes were synthesized through the hydroboration of corresponding

polysilazanes. The conditions used were the same as in the case of polyborophenylvinyl-

silylcarbodiimide and polyboromethylvinylsilylcarbodiimide. The reactions for forming B[-

22 Experimental

(C2H4)Si(Ph)-NH-]3n and B[-(C2H4)Si(Me)-NH-]3n are shown in Scheme 10 and Scheme

11, respectively.

Scheme 10 Synthesis of polyborophenylvinylsilazane

Scheme 11 Synthesis of polyboromethylvinylsilazane

3.2.5 Synthesis of polyphenylsilsesquicarbodiimides

A branched polyphenylsilsesquicarbodiimide was synthesized using

bis(trimethylsilyl)carbodiimide and trichlorophenylsilane. Pyridine was used as the catalyst.

The reaction is shown in Scheme 12.

Scheme 12 Synthesis of polyphenylsilsesquicarbodiimide

13C/15N isotope-enriched polyphenylsilsesquicarbodiimide was synthesized from

trichlorophenylsilane and cyanamide with pyridine as the catalyst. 10 % of the cyanamide was

labeled with 15N and 13C isotopes. Cyanamide and pyridine were mixed with the solvent THF.

The mixture was first stirred in an ice bath for one hour. Trichlorophenylsilane was diluted by

THF and then dropped into the mixture of cyanamide and pyridine. After the dropping

finished, the mixture was further stirred for one hour in the ice bath. Then, the ice bath was

removed and the mixture was warmed up to room temperature and kept under stirring for 12

hours. After removing the THF by vacuum drying, the by-product, pyridine·HCl, was

removed by washing the polymer with pyridine. More THF was used to wash away the

residual pyridine, and at last, THF was removed by performing vacuum dring. The 15N and 13C isotopes go into the carbodiimide groups forming the structure [-(Ph)Si-(15N=13C=15N)1.5-

]n. The reaction is shown in Scheme 13.

Experimental 23

Scheme 13 Synthesis of 13C/15N-isotopes enriched polyphenylsilsesquicarbodiimide

3.2.6 Synthesis of polyphenylsilsesquiazane

Polyphenylsilsesquiazane was synthesized from the reaction of hexamethyldisilazane

and trichlorophenylsilane with pyridine as catalyst. The reaction was carried out at room

temperature for 0.5 hour, and then at 120C for another 10.5 hours. The by-product,

trimethylchlorosilane, was removed by vacuum drying. The final state of the polymer is a

white powder. The reaction is shown in Scheme 14.

Scheme 14 Synthesis of polyphenylsilsesquiazane

3.3 Pyrolysis of polymers

The samples studied in this work included both ceramic powders and bulk ceramics.

Ceramic powders were obtained from a direct pyrolysis of the polymers, while bulk ceramics

were synthesized using the procedure described in Section 3.4.

To obtain ceramic powders, the precursors were pyrolyzed in a horizontal tube furnace

(Gero GmbH & Co., Neuhausen, Germany). The polymer was put into a quartz boat, which

was kept in a quartz Schlenk tube. The polymers were pyrolyzed at a heating rate of 100 °C/h

to the target temperature under constant argon flow. At the desired temperature, the material

was held for 2 hours. The furnace was then cooled to room temperature at a cooling rate of

100 °C/h to 700°C, followed by furnace cooling. The yields of ceramics at 1100°C are given

in Table 1.

The green bodies of bulk ceramic were put in a SiC boat, and kept in quartz Schlenk

tube. The pyrolysis schedule differed with the precursors according to their mass loss features

— details can be found in Section 3.4.

The pyrolyzed ceramic samples are identified hereon as “Si(B)CNX-pyrolysis

temperature,” in which X represents a sequence number given in Table 1. For instance,

SiBCN4-1400 stands for the ceramic sample derived from polyboromethylvinylsilazane

pyrolyzed at 1400C.

24 Experimental

Table 1 Ceramic yields and identification

Preceramic polymer Polymer formula Ceramic

yield (1100°C) Ceramic sample

polyphenylvinylsilylcarbodiimide [-(PhVi)Si-NCN-]n 61.5% SiCN1

polymethylvinylsilylcarbodiimide [-(MeVi)Si-NCN-]n 32.7% SiCN2

polyborophenylvinylsilylcarbodiimide B[-(C2H4)Si(Ph)-NCN-]3n 59.3% SiBCN1

polyboromethylvinylsilylcarbodiimide B[-(C2H4)Si(Me)-NCN-]3n 55.7% SiBCN2

polyphenylvinylsilazane [-(PhVi)Si-NH-]n 59.1% SiCN3

polymethylvinylsilazane [-(MeVi)Si-NH-]n 35.3% SiCN4

polyborophenylvinylsilazane B[-(C2H4)Si(Ph)-NH-]3n 46.6% SiBCN3

polyboromethylvinylsilazane B[-(C2H4)Si(Me)-NH-]3n 35.6% SiBCN4

polyphenylsilsesquicarbodiimide [-(Ph)Si-(NCN)1.5-]n — SiCN5 13C/15N isotope-enriched

polyphenylsilsesquicarbodiimide

13C/15N isotope-enriched [-(Ph)Si-(NCN)1.5-]n

48.1% Iso_SiCN5

polyphenylsilsesquiazane [-(Ph)Si-(NH)1.5-]n 49.3% SiCN6

3.4 Bulk ceramic processing

Bulk ceramics were produced from all linear polymers. Figure 9 shows the processing

route for producing bulk ceramics, together with that for preparing ceramic powders. The bulk

ceramics were synthesized using a uniaxial warm-pressing, followed by pyrolysis of the green

bodies. All liquid (honey-like) polymers, namely [-(PhVi)Si-NCN-]n, [-(MeVi)Si-NCN-]n, [-

(PhVi)Si-NH-]n, [-(MeVi)Si-NH-]n, required a pre-crosslinking step. The required

temperatures and dwelling times varied, as shown in Table 2. After pre-crosslinking, the

products obtained were milled into a powder, followed by warm-pressing. The parameters

used for warm-pressing are shown in Table 2. Compacts and mechanically-stable green

bodies were then produced.

Figure 9 The main procedure utilized for the synthesis of ceramic powder and bulk ceramic

warm press

pyrolysis

pre‐crosslinking

mixed with original polymer, warm press

pre‐pyrolysis

pyrolysis

hydroboration pyrolysis

Ceramic Powder

Bulk Ceramic

Ceramic Powder

Green body Bulk Ceramic

pre‐ceramic self‐filler Green body pyrolysis

Bulk Ceramic

warm press pyrolysis

Polyborosilazanes

Polysilazanes

Polysilylcarbodiimides

Polyborosilylcarbodiimides

Green body partially

crosslinked

polymer

Experimental 25

Polyborophenylvinylsilazane and polyboromethylvinylsilazane are thermoplastic with

similar melting points around 85°C. In order to obtain bulk samples, the polymers were first

pyrolyzed at 500 °C for 2 hours. The pre-pyrolyzed powder and the original polymer were

then mixed for warm-pressing. The optimized ratio of pre-pyrolysis powder to polymer was

65:35 for B[-(C2H4)Si(Ph)-NH-]3n, and 50:50 for B[-(C2H4)Si(Me)-NH-]3n.

The processing parameters, including pre-crosslinking/pre-pyrolysis temperature and

dwelling time, warm-pressing temperature, pressing time, and applied pressure are listed in

Table 2.

Table 2 Optimized parameters for processing bulk ceramics

Preceramic polymer

Pre-crosslink/pre-

pyrolysis Warm-pressing

Pyrolysis schedule temperature,

time temperature, holding

time, pressure

[-(PhVi)Si-NCN-]n 255C, 5h 250C, 2h, 16MPa

r.t. (100C/h)400C (50C/h)500C (2h)500C (50C/h)600C (2h)600C (100C/h)1100C (2h)1100C

[-(MeVi)Si-NCN-]n 320C, 5h 200C, 2h, 16MPa

r.t. (100C/h)450C (50C/h)500C (2h)500C (50C/h)700C (100C/h)1100C (2h)1100C

B[-(C2H4)Si(Ph)-NCN-]3n — 250C, 2h, 16MPa r.t. (100C/h)200C (50C/h)600C (100C/h)1100C (2h)1100C

B[-(C2H4)Si(Me)-NCN-]3n — 250C, 2h, 16MPa r.t. (100C/h)300C (50C/h)700C (100C/h)1100C (2h)1100C

[-(PhVi)Si-NH-]n 300C, 5h 250C, 2h, 16MPa

r.t. (100C/h)200C (50C/h) 300C (1h)300C (50C/h) 450C (1h)450C (50C/h) 600C (100C/h)1100C (2h)1100C

[-(MeVi)Si-NH-]n 320C, 5h 300C, 2h, 16MPa

r.t. (50C/h)160C (2h)160C (50C/h) 450C (1h)450C (50C/h) 600C (100C/h)1100C (2h)1100C

B[-(C2H4)Si(Ph)-NH-]3n 500C, 2h

85C, 1.5h, 16MPa (pre-pyrolyzed

powder: original polymer = 65:35 by

weight)

r.t. (100C/h)200C (25C/h)600C (100C/h)1100C(2h)1100C

B[-(C2H4)Si(Me)-NH-]3n 500C, 2h

85C, 1.5h, 16MPa (pre-pyrolyzed

powder: original polymer = 50:50 by

weight)

r.t. (100C/h)200C (25C/h) 300C (1h)300C (25C/h) 400C (1h)400C (25C/h) 500C(1h)500C (25C/h) 600C(1h)600C (100C/h)1100C (2h)1100C

26 Experimental

3.5 Annealing of ceramics

The post-annealing experiments at 1400 °C, 1700 °C and 2000 °C were performed in

an Astro furnace (Thermal Technology Inc., CA, USA) in argon atmosphere. Ceramics

obtained at 1100 °C were placed in a silicon carbide crucible and pyrolyzed at a heating ramp

of 5 °C/min to target temperatures, and held for 2 hours. The thermolysis was completed by

cooling the samples to room temperature at a cooling ramp of 10 °C/min.

3.6 Characterization techniques

Micro-Raman spectra (10 scans, each lasting 3s) were recorded using a Horiba HR800

micro-Raman spectrometer (Horiba Jobin Yvon, Bensheim, Germany) equipped with an Ar

laser (irradiation wavelength of 514.5nm). The excitation line has its own interference filter

(to filter out the plasma emission) and a Raman notch filter (for laser light rejection). The

measurements were performed with a grating of 1800 gmm−1 and a confocal microscope

(magnification 50×, NA0.5) with a 100 μm aperture, giving a resolution of 2–4 μm. The laser

power (ca. 20 mW) on the sample was attenuated in the range of 2mW–20μW using neutral

density (ND) filters. For the evaluation of graphitic carbon in ceramics, Gaussian/Lorentz

curve fitting of the Raman bands (OriginPro 8.1 Software) was applied.

All-reflecting objective (ARO) IR micro-spectroscopy was done on a confocal

Illuminat IR™ FTIR spectrometer using reflection absorption spectroscopy (RAS) for quick

analysis. The spectrum of the microscopic specimen is recorded by simply focusing on the

specimen.

Scanning electron microscopy (SEM) micrographs were obtained using an FEI Quanta

600 instrument (FEI, Eindhoven, The Netherlands). The samples were sputtered with a

conductive gold layer prior to investigation.

For the chemical composition analysis, the carbon content of the samples was

determined with a carbon analyzer, CS 800 (Eltra GmbH, Neuss) while the oxygen and

nitrogen content was measured with a N/O analyzer, Leco TC-436 (Leco Corporation,

Michigan). Hydrogen, boron and chlorine contents of the ceramic pyrolyzed at 1100 °C were

carried out at Mikroanalytisches Labor Pascher (Remagen, Germany). The silicon content of

the samples was calculated by mass balance.

Transmission electron microscopy (TEM) of the bulk and powder ceramics was

performed using an FEI CM20 STEM (FEI, Eindhoven, the Netherlands) operating at an

accelerating voltage of 200 kV (wavelength λ=2.51 pm). For the TEM sample preparation,

ceramic powders were dispersed in an ultrasonic bath of high purity (99.8%) methanol

(Sigma-Aldrich Co.), and a small droplet of the suspension was drop-casted onto a holey

carbon (Cu) grid. Upon drying, the samples were lightly coated with carbon to avoid charging

under the electron beam. TEM samples of the bulk materials were prepared via conventional

Experimental 27

preparation techniques involving cutting, grinding and polishing, followed by Ar-ion milling

and light carbon coating.

Powder X-ray diffraction (XRD) measurements were performed on a STOE X-ray

diffractometer using Ni-filtered Mo K radiation at a scan speed of 1º min-1. For the bulk

ceramics, XRD measurements were performed on a modified 4 circle X-ray diffractometer

(STOE STADI 4). The system was equipped with new FOX X-ray optics with a curved

mirror with graded multilayer coating from XENOCS, and a collimator with a divergence of

0.06° primary to the samples. A Cu K anode (Seifert) was applied. The step size was 0.05°,

and each step was measured for 60 or 90 seconds. The surface of the bulk ceramics were

polished prior to the measurements.

Thermogravimetry coupled with mass spectrometry (TG/MS) was used to characterize

the pyrolysis behavior of the polymers. The TGA signal was recorded using a Netzsch STA

449 F1 Jupiter. Gases evolved during thermolysis were analyzed by a quadrupole mass

spectrometer (Netzsch QMS 403 C Aeolos). The polymers were heated under an argon

atmosphere from room temperature to 1400 ºC at a rate of 100 ºC/h.

High temperature TG was carried out at UC Davis using a Setaram SetSys 2400 in

helium atmosphere. About 20-30 mg of powders previously pyrolyzed at 1100 ºC was placed

into a lidded graphite crucible. The samples were heated at 10 °C/min from 1100 to 2100 °C.

After the first run of each sample, a second, identical run was completed in order to obtain a

background signal that was subtracted from the initial raw data.

Nuclear magnetic resonance (NMR) spectroscopy was conducted at room temperature

on liquid state polymers using a 500 MHz Bruker DRX 500 spectrometer. The chemical

environments of 1H, 13C and 29Si were characterized. Tetramethylsilane was used as an

external reference. Measurements were performed on 0.4 ml of sample mixed with 0.2 ml of

deuterated C6D6 solvent.

Multinuclear MAS NMR spectra were collected at UC Davis on SiCN1-800, SiCN1-

1100, SiBCN1-1100, SiBCN3-1100, SiCN4-1100, SiCN4-1400, SiBCN4-1100, SiBCN4-

1400, Iso_SiCN5-800, and Iso_SiCN5-1100 using a Bruker Avance solid-state spectrometer

operating at the Larmor frequencies of 99.3, 125.7, and 160.4 MHz, respectively, for the 29Si, 13C and 11B nuclides. The ceramics were ground and then loaded into ZrO2 rotors under argon

atmosphere. Samples were spun at a rate of 14 kHz using a 4 mm Bruker CPMAS probe. All

NMR spectra were externally referenced to tetramethylsilane for 29Si and 13C, and to a 0.1 M

Na2B4O7 solution for 11B. Specific experimental parameters have been reported

previously [131].

Solid state MAS NMR was performed on [-(PhVi)Si-NH-]n pre-crosslinked material,

[-(MeVi)Si-NH-]n pre-crosslinked material; B[-(C2H4)Si(Ph)-NH-]3n and B[-

(C2H4)Si(Me)-NH-]3n polymers; SiCN1-1400, SiBCN1-1400, SiCN2-1100, SiCN2-1400,

SiBCN2-1100, SiBCN2-1400, SiCN3-1100, SiCN3-1400 and SiBCN3-1400 ceramics. The

28 Experimental

device used was a house-built 11.7 Tesla spectrometer with a 4 mm rotor probehead and with

the frequencies of proton at 500.144 MHz, carbon at 125.777 MHz, and silicon at 99.355

MHz. The samples were spun at 6 kHz.

Small Angle X-ray scattering (SAXS) measurements on SiCN1 and SiBCN1 ceramics

were completed using a SAXS device equipped with a Cu anode operating at 40 kv and 55

mA. Other equipment details include a 3-pinhole point focus, a 2-dimensional detector, and a

designed q-range of 0.08 – 2.8 nm-1. The usable q-range is between 0.014 and 0.5 1 Å-1.

Measurements were taken at room temperature. Samples were affixed to the sample chamber

using office tape, and no sample holder was used.

SAXS measurements on SiCN2, SiBCN2, SiCN3, SiBCN3, SiCN4 and SiBCN4

ceramics prepared at 1100C and 1400C were measured at the Hamburger

Synchrotronstrahlungslabor (HASYLAB) at the Deutsches Elektronen-Synchrotron (DESY)

Research Centre of the Helmholtz Association, in Hamburg, Germany. Beamline A2 of the

DORIS III storage ring was used. The wavelength was 0.15 nm, and the sample-detector

distance was 1200 mm. The samples were loaded into 0.5 mm capillaries.

High temperature oxidative drop solution calorimetry in molten sodium molybdate

solvent, as previously described in detail by Navrotsky [92,93], was conducted at UC Davis

in order to determine the enthalpies of structural evolution. Pressed powder pellets (1-2 mg)

were dropped from room temperature into molten sodium molybdate (3Na2O·4MoO3) solvent

at about 800 C in a custom-built Tian-Calvet twin microcalorimeter (Davis, CA). In order to

maintain a oxidizing atmosphere, and to encourage rapid sample dissolution by solvent

agitation, oxygen gas was bubbled through the solvent at 5 ml·min-1. All pellets dissolved

within 1 hour, and were converted to dissolved oxides with the evolution of CO2, N2, and H2O

gases. To remove the evolved gases, and to maintain a constant atmosphere, oxygen gas was

used to flush the headspace above the solvent at a rate of 90 ml·min-1. At least 8 experiments

were carried out for each sample in order to obtain acceptable statistics. To determine the

enthalpies of drop solution, ΔHds, the calorimeter was calibrated using the heat capacity of

platinum. Additional details regarding calorimeter calibration can be found in the

literature [95,132].

Impedance spectroscopy measurements were taken using a Broadband

Dielectric/Impedance Spectrometer from Novocontrol Technologies. The system was

equipped with a Alpha-A high performance frequency analyzer. Frequencies from 0.1Hz to

107 Hz were applied. The ceramic samples were polished and covered with silver paste prior

to the measurements. Simulations of the results were carried out using Zview2 Software.

The ceramics used for the electrochemical studies were ground and sieved using a 40

µm mesh. For the electrode preparation, a mixture of 85 wt% of the ceramic, 5 wt% of

Carbon Black Super P (Timcal Ltd., Switzerland) and 10 wt% of polyvinylidenefluoride

(PVdF, SOLEF, Germany) was dissolved into N-methyl-2-pyrrolidone (NMP, BASF,

Experimental 29

Germany). Additional NMP was added to adjust the viscosity of the mixture. The slurry was

printed onto the rough side of a copper foil (10 µm, Copper SE-Cu58 (C103), Schlenk

Metallfolien, Germany) by hand, and dried at 80°C for 24 h. Electrodes (7mm diameter) were

cut from the coated copper foil and dried at 80°C under vacuum in a Buchi oven for 24 h. The

dried electrodes were transferred to a glove box (MBraun, Germany) without further contact

to air for the cell assembly. For electrochemical testing a Swagelok type cell set up was

chosen. The counter/reference electrode was cut out of metallic lithium foil (99.9% purity,

0.75 mm thickness, Alfa Aesar, Germany) to have a diameter of 10 mm. QMA (WhatmannTM,

UK) was used as a separator and 180 µl of LP30 (Merck, Germany) was added as the

electrolyte. Electrochemical testing was performed using a VMP multipotentiostat (BioLogic

Science Instruments). The same rate was used for the charge (C) and discharge (D) processes,

with C = D at C/20 = 18 mA·g-1, C/10 = 36 mA·g-1, C/5 = 72 mA·g-1, C/ 2 = 180 mA·g-1, and

C/1 = 360 mA·g-1. Charge is considered to be lithium insertion, while discharge is considered

to be lithium extraction. All capacity calculations were completed taking into account that the

active material only made up 85 wt % of the electrode.

3.7 Characterization of ceramics prepared at selected temperatures

Several pyrolysis temperatures have been chosen. 800ºC is the temperature at which

the precursor begins to transform from polymer to ceramic, thus considerable hydrogen

remains in the structure. 1100C was selected since it is the standard temperature for

producing amorphous PDCs. In most SiCN and SiBCN PDCs, 1400C is considered the

temperature at which phase separation begins, but no deleterious crystallization initiates.

Table 3 gives the chemical compositions of the ceramics. The chemical formulas were

calculated by considering one Si atom per formula unit. The stoichiometric binary phases

were also obtained based on elemental content.

Experimental 30

Table 3 Chemical compositions of the synthesized ceramics

Preceramic polymerCeramic sample

Si wt%

atom%

C wt%

atom%

N wt%

atom%

B wt%

atom%

O wt%

atom%

H wt%

atom%

Cl wt%

atom%Empirical formula

Calculated phases (Mol%)

Free C

Si3N4 SiC BN SiO2

[-(PhVi)Si-NCN-]n

SiCN1-800 26.10 12.14

54.61 59.40

15.93 14.88

— 2.46 1.96

0.89 11.62

— SiC4.88N1.22O0.16 92.61 5.79 0.04 — 1.56

SiCN1-1100 24.80 12.53

56.50 66.61

14.80 14.95

— 2.14 1.89

0.24 3.40

1.55 0.62

SiC5.32N1.19O0.15 92.91 5.24 0.52 — 1.33

SiCN1-1400 29.24 15.36

59.21 72.57

11.03 11.59

— 0.52 0.48

— — SiC4.72N0.75O0.03 87.37 3.83 8.49 — 0.32

B[-(C2H4)Si(Ph)-NCN-]3n

SiBCN1-1100 23.79 12.62

51.32 63.54

16.58 17.59

3.05 4.20

2.21 2.05

— — SiC5.03N1.39B0.33O0.16 85.97 4.65 2.15 5.82 1.42

SiBCN1-1400 28.21 14.83

51.08 62.64

16.19 17.02

3.06 4.17

1.46 1.34

— — SiC4.22N1.15B0.28O0.09 82.22 4.54 6.39 5.90 0.95

[-(MeVi)Si-NCN-]n

SiCN2-1100 39.95 23.49

32.37 44.42

24.53 28.85

— 3.15 3.24

— — SiC1.89N1.23O0.14 82.97 13.54 0.44 — 3.04

SiCN2-1400 40.05 23.51

33.65 46.09

22.98 26.98

— 3.32 3.41

— — SiC1.96N1.15O0.14 81.63 12.37 2.88 — 3.13

B[-(C2H4)Si(Me)-NCN-]3n

SiBCN2-1100 37.71 21.78

30.41 40.98

24.67 28.50

3.00* 4.49*

4.21 4.25

— — SiC1.88N1.31B0.21O0.20 73.38 11.20 3.08 8.38 3.97

SiBCN2-1400 39.94 23.34

30.31 41.33

23.43 27.39

3.00* 4.55*

3.32 3.40

— — SiC1.77N1.17B0.19O0.15 69.10 10.72 8.47 8.53 3.19

[-(PhVi)Si-NH-]n SiCN3-1100

31.31 16.53

52.52 64.68

13.01 13.73

— 2.41 2.23

0,18 2.59

0.57 0.24

SiC3.91N0.83O0.13 86.0 5.0 7.4 — 1.6

SiCN3-1400 33.23 18.05

51.83 65.71

13.76 14.95

— 0.64 0.61

0.03 0.46

0.51 0.22

SiC3.64N0.83O0.03 84.8 5.4 9.4 — 0.4

B[-(C2H4)Si(Ph)-NH-]3n

SiBCN3-1100 27.71 14.28

55.18 66.37

11.85 12.22

2,57 3.37

2.27 2.05

0,11 1.59

0.305 0.12

SiC4.65N0.86B0.24O0.14 81.9 3.0 9.1 4.6 1.4

SiBCN3-1400 27.77 14.49

55.44 67.50

12.20 12.52

2.55 3.39

1.94 1.77

0.02 0.22

0.28 0.12

SiC4.66N0.86B0.23O0.12 82.0 3.1 9.1 4.6 1.2

Experimental 31

* samples have not been measured, the values are estimated from the ratio of starters in polymer reactions and the ceramic yield.


Si wt%

atom%

C wt%

atom%

N wt%

atom%

B wt%

atom%

O wt%

atom%

H wt%

atom%

Cl wt%



Free C

Si3N4 SiC BN SiO2

[-(MeVi)Si-NH-]n

SiCN4-1100 51.75 31.01

29.19 40.81

17.10 20.49

— 1.56 1.64

0.36 6.04

0.04 0.02

SiC1.32N0.66O0.05 55.6 11.0 31.7 — 1.8

SiCN4-1400 51.283 31.55

29.84 42.84

17.23 21.20

— 1.44 1.55

0.165 2.84

0.042 0.02

SiC1.36N0.67O0.05 57.2 10.8 30.4 — 1.6

B[-(C2H4)Si(Me)-NH-]3n

SiBCN4-1100 48.36 29.70

28.85 41.35

13.60 16.71

4.52 7.07

4.63 4.98

0.01 0.17

0.03 0.01

SiC1.39N0.56B0.24O0.17 40.1 4.5 37.5 13.3 4.7

SiBCN4-1400 48.20 29.55

28.39 40.62

14.09 17.28

4.68 7.44

4.59 4.93

0.01 0.17

0.04 0.02

SiC1.37N0.58B0.25O0.17 39.5 4.6 37.2 14.0 4.7

13C/15N isotope-enriched [-Si(Ph)-

(NCN)1.5-]n

Iso_SiCN5-800 24.03 9.59

37.08 34.53

30.54 24.38

— 5.90 4.12

2.45 27.38

— SiC3.60N2.54O0.43 74.5 7.9 C3N4

11.2— 6.5

Iso_SiCN5-1100 28.11 13.39

39.16 43.52

29.94 28.52

— 1.81 1.51

0.98 13.07

— SiC3.25N2.13O0.11 81.5 9.9 C3N4

6.8 — 1.8

[-Si(Ph)-(NH)1.5-]n SiCN6-1100 34.59 18.43

46.86 58.25

16.04 17.09

— 2.13 0.99

0.35 5.22

0.03 0.01

SiC3.16 N0.93O0.11 84.4 6.7 7.3 — 1.6

32 Experimental

Figure 10 reveals that in a composition diagram, the compositions of all ceramics

pyrolyzed at 1100 ºC lie within the triangular region constituted by Si3N4, SiC and “free”

carbon. The compositions of SiCN1-1100, SiBCN1-1100, SiCN2-1100 and SiBCN2-1100

ceramics, which are derived from poly(boro)silylcarbodiimides, are near the Si3N4-carbon tie-

line. Meanwhile, ceramics derived from poly(boro)silazanes, by calculation, have a greater

SiC component.

0,00 0,25 0,50 0,75 1,00

0,00

0,25

0,50

0,75

1,000,00

0,25

0,50

0,75

1,00

SiCN6

Iso_SiCN5

SiCN1SiBCN1

SiCN2

SiBCN3

SiBCN4

SiCN3

SiCN4

C

NSi Si3N4

SiC

SiBCN2

Figure 10 The compositions of the ceramics pyrolyzed at 1100 ºC marked in a Si-C-N ternary composition

diagram

Disregarding hydrogen and boron contents, the compositions of Si, C, N elements in

the synthesized poly(boro)phenylvinylsilylcarbodiimide, poly(boro)methylvinylsilylcarbo-

diimide, poly(boro)phenylvinylsilazane, poly(boro)methylvinylsilazane and 13C/ 15N isotope-

enriched polyphenylsilsesquicarbodiimide have the formulas SiN2C9, SiN2C4, SiNC8, SiNC3,

SiN3C7.5 and SiN1.5C6, respectively. From the locations of their compositions in the diagram,

Si atoms prefer to bond to N to form Si3N4, and extra Si bonds to C to form SiC. In the

Iso_SiCN5 ceramics, which contain considerably more N than the other ceramics, after

bonding to Si, residual N bonds to C. The existence of “free” carbon is a common

characteristic in all of these ceramics.

Polymer synthesis and processing 33

4. Results and discussion

4.1 Polymer synthesis and processing

The procedures for polymer synthesis were given in Chapter 3. In this Section,

important characterization results for the obtained polymers are presented. The critical

processing steps for producing bulk ceramics, namely polysilylcarbodiimides and

polysilazanes pre-crosslinking and polyborosilazanes pre-pyrolysis, are discussed in Sections

4.1.1 and 4.1.2, respectively. The formation of branched polysilylsesquicarbodiimide is

discussed in Section 4.1.3, followed by the synthesis of its 15N/ 13C isotope-enriched

counterpart. More details of the polymer synthesis can be found in the references [131,133–

137].

4.1.1 Synthesis of poly(boro)silylcarbodiimides and bulk ceramic processing

The polymer precursors were characterized by NMR spectroscopy. Figure 11(a) shows

the 29Si NMR DEPT spectra for both polyphenylvinylsilylcarbodiimide and

polymethylvinylsilylcarbodiimide after by-product removal. The signal at -46 ppm represents

the Si environment in -(NCN)-(PhVi)Si-(NCN)- and the peak at -34 ppm is from the end

group -(NCN)-(PhVi)Si-Cl. The strong peak at -46 ppm indicates that the linear [-(PhVi)Si-

NCN-]n is obtained through the designed route. The reaction products of [-(MeVi)Si-NCN-]n

are shown by three significant peaks. A shift at -37 ppm is from the basic polymer structure, -

(NCN)-(MeVi)Si-(NCN)-. This result verifies that [-(MeVi)Si-NCN-]n was successfully

synthesized. Two kinds of end groups, -SiMe3 and -(NCN)-(MeVi)Si-Cl, are found located at

ca. 0 ppm and -12 ppm, respectively.

Figure 11 (b) compares the 1H NMR spectra of [-(PhVi)Si-NCN-]n and [-(MeVi)Si-

NCN-]n. [-(PhVi)Si-NCN-]n shows peaks between 7 ppm and 8 ppm, which is typical for

phenyl groups. The multiple peaks between 5.5 ppm and 6 ppm in both polymers result from

vinyl groups. The polymer, [-(MeVi)Si-NCN-]n, shows strong peaks around 0 ppm due to the

methyl groups—the functional methyl groups, and the methyl groups from the end-group -

SiMe3. [-(PhVi)Si-NCN-]n also has methyl groups which are end-groups in the same chemical

shift range. There are two additional small peaks in the [-(PhVi)Si-NCN-]n spectra at ca. 0.9

ppm and 1.5 ppm, which represent the -CH2- and -CH- groups. These signals may be from the

small amount of saturated vinyl groups from the polymer synthesis.

Figure 11(c) shows the 13C NMR spectra of [-(PhVi)Si-NCN-]n and [-(MeVi)Si-NCN-

]n. The peaks belonging to the -N=C=N- groups and the benzene solvent C6D6 are at ca. 120

ppm. Phenyl groups have chemical shifts between 126 ppm and 131 ppm. Signals for the

vinyl groups are located at about 136 ppm. In addition to these peaks, [-(MeVi)Si-NCN-]n is

also characterized by the peak around 0 ppm, which is again the signal for the methyl

substituents and methyl end-groups. The 13C NMR DEPT spectra (Figure 11 (d)) differ

between the odd and even protons bonded to carbon atoms. The signals from the -N=C=N-

34 Polymer synthesis and processing

groups and C6D6 are not detectable in the 13C DEPT NMR spectra. The difference between

the carbons in the phenyl and vinyl groups can be distinguished with the help of 13C NMR.

Carbon atoms with an even number of protons have negative intensities, as shown by the

signal at 136 ppm for the vinyl group in Figure 11(d). Signals from phenyl groups have

positive intensity.

40 20 0 -20 -40 -60

[-(MeVi)Si-NCN-]n

[-(PhVi)Si-NCN-]n

29Si chemical shift (ppm)

10 9 8 7 6 5 4 3 2 1 0 -1

[-(MeVi)Si-NCN-]n

[-(PhVi)Si-NCN-]n

1H chemical shift (ppm)

200 150 100 50 0

[-(MeVi)Si-NCN-]n

[-(PhVi)Si-NCN-]n

13C chemical shift (ppm)

200 150 100 50 0 -50

[-(MeVi)Si-NCN-]n

[-(PhVi)Si-NCN-]n

13C DEPT chemical shift (ppm)

Figure 11 (a) 29Si, (b)1H, (c)13C, (d)13C DEPT NMR spectra of [-(PhVi)Si-NCN-]n and [-(MeVi)Si-NCN-]n

As previously discussed, the 29Si NMR spectra show that [-(PhVi)Si-NCN-]n and [-(MeVi)Si-NCN-]n are linear polymers. Since they are liquid (honey-like) after synthesis, in order to accomplish bulk ceramic shaping, a pre-crosslinking process is needed. This is the most critical step, since polymer solidification is needed through the polymerization of vinyl groups. Meanwhile, some of the C=C bonds must remain active for the warm pressing step (Figure 12, [-(PhVi)Si-NCN-]n as an example). Partial pre-crosslinking, as opposed to full crosslinking, is essential. This step was optimized after a systematic study of varied temperatures and heat treatment times.

Figure 12 partially pre-crosslinking step of [-(PhVi)Si-NCN-]n

Si N NCn pre crosslinking

in oil bath

1Si N NC

xSi N NC

yn

R

warm press

2Si N NC

n

R

R= SiPhNCN R= SiPhNCN

(a) (b)

(c) (d)


For [-(PhVi)Si-NCN-]n, heat treating at 255°C for 5 h is optimal, while for [-

(MeVi)Si-NCN-]n, 320°C for 5 h is optimal. From 1H NMR (Figure 13), some vinyl groups

remain after pre-crosslinking, and the signal intensity of the saturated -CH2- and -CH-

increased. The partially saturated vinyl groups meet the needs of warm-pressing.

8 7 6 5 4 3 2 1 0

[-(PhVi)Si-NCN-]n

pre-crosslinked [-(PhVi)Si-NCN-]n

-C-H=C-H

=C-H

-C-H


8 7 6 5 4 3 2 1 0

[-(MeVi)Si-NCN-]n

-C-H


=C-H

=C-H

pre-crosslinked [-(MeVi)Si-NCN-]n

Figure 13 1H NMR of the comparison of polymer and pre-crosslinked material for (a) [-(PhVi)Si-NCN-]n and (b)

[-(MeVi)Si-NCN-]n

Boron-modified polyborophenylvinylsilylcarbodiimide and polyboromethylvinylsilyl-

carbodiimide are synthesized via hydroboration of polyphenylvinylsilylcarbodiimide and

polymethylvinylsilylcarbodiimide, respectively. Vinyl groups were mostly saturated by

reaction with BH3. The obtained polyborophenylvinylsilylcarbodiimide and polyboromethyl-

vinylsilylcarbodiimide have crosslinked network structures. No extra pre-crosslinking process

is needed for warm-pressing.

Green body compacts were produced via the warm presssing of polyborosilylcarbo-

diimides and pre-crosslinked polysilylcarbodiimides. Figure 14 shows the warm-pressed

green bodies.

Figure 14 Warm pressed greenbodies derived from (a) [-(PhVi)Si-NCN-]n, (b) [-(MeVi)Si-NCN-]n, (c) B[-

(C2H4)Si(Ph)-NCN-]3n, (d) B[-(C2H4)Si(Me)-NCN-]3n

4.1.2 Synthesis of poly(boro)silazanes and bulk ceramic processing

Polysilazanes as polymer precursors were traditionally sythesized by ammonolysis or

aminolysis reactions of chlorosilane [56]. Polyborosilazanes are usually synthesized through

(a) (b) (c) (d)

(a) (b)


the hydroboration of vinyl-containing polysilazanes. In this reaction, borane reacts with the

vinyl groups, and the saturation of the vinyl groups leads to cross-linking and the formation of

a 3D network structure.

The newly obtained polysilazane polymers were synthesized using a reflux method.

Polyborosilazanes are the hydroboration products of polysilazanes. The synthesis route is

described in Chapter 3. The resultant polyborosilazanes are solid at room temperature, but

can be melted at temperatures greater than 85°C. This offers a good opportunity to employ

various plastic forming techniques to produce polymer-derived SiBCN ceramic parts.

Another advantage of this reflux synthesis route of polysilazanes is that it avoids the

filtration of NH4Cl, which was formed as a solid by-product, and is not easily separated from

the mixture in the traditional synthesis of polysilazanes. In this work, the by-product, which

was formed during the synthesis of [-(PhVi)Si-NH-]n and [-(MeVi)Si-NH-]n, namely Me3SiCl,

was liquid and could be easily removed by vacuum drying. There also have been studies on

using a reflux method for the synthesis of polysilazane, without a catalyst [69,70], or with

AlCl3 as the catalyst [70]. However, the properties of the polymers produced by these

methods have not been reported.

29Si and 13C DEPT NMR measurements were completed on the as-synthesized

polysilazanes (Figure 15). Two bands at ca. -25 ppm for [-(PhVi)Si-NH-]n (Figure 15(a))

indicate the Si environment in the -NH-Si(PhVi)-NH- group. The peaks between 0 and 10ppm

are from the intermediate product, Cl-(PhVi)Si-NH-SiMe3, and the residual monomer,

phenylvinyldichlorosilane, respectively. The residual by-product, Me3SiCl, shows a weak

peak at 30ppm. The spectrum of [-(MeVi)Si-NH-]n shows peaks at ca.-12 to -22ppm which

indicate -NH-(MeVi)Si-NH-. The signal for the end group, -(MeVi)Si-Cl, appears at ca. -

10ppm. A signal at 2ppm is the residual starter HMDS. The 13C DEPT NMR spectrum

(Figure 15(b)) contains two peaks between 130 ppm and 140 ppm corresponding to =CH- and

=CH2, as observed for both [-(PhVi)Si-NH-]n and [-(MeVi)Si-NH-]n. In addition, [-(PhVi)Si-

NH-]n also has a signal at 125 ppm, which represents phenyl groups. These spectra confirm

the formation of expected polymer structures.


40 20 0 -20 -40

[-(MeVi)Si-NH-]n

[-(PhVi)Si-NH-]n


250 200 150 100 50 0 -50

[-(MeVi)Si-NH-]n

[-(PhVi)Si-NH-]n


CH2

CH2

CH3

CH3

CH2

CH2

CH

CH

Figure 15 (a) 29Si DEPT NMR (b) 13C DEPT NMR spectra of [-(PhVi)Si-NH-]n and [-(MeVi)Si-NH-]n

Similar to polysilylcarbodiimides, polysilazanes need a pre-crosslinking process (by

heating) before warm pressing. The optimized heating temperature and treatment time are

300ºC for 5 hours for [-(PhVi)Si-NH-]n, and 320ºC for 5 hours for [-(MeVi)Si-NH-]n. 29Si and

13C MAS NMR were measured on pre-crosslinked polymers. The 29Si MAS NMR of pre-

crosslinked [-(PhVi)Si-NH-]n (Figure 16(a)) mainly has two peaks around -26 and -15 ppm,

which represent -N-(PhVi)Si-N- cites and -N-Si(PhCsp3)-N- sites. Sp3 C in -N-Si(PhCsp3)-N-

(a)

(b)


sites are from the polymerisation of vinyl groups. The bands of [-(MeVi)Si-NH-]n are mainly

centered at -18 and -5 ppm signify -N-(MeVi)Si-N- cites and -N-Si(MeCsp3)-N- sites. The

positions of these Si environments are reported in reference [102]. Similarly, sp3 C in -N-

Si(MeCsp3)-N- sites are from vinyl saturation. These results demonstrate that vinyl groups are

partially saturated in the pre-crosslinking step. The unsaturated vinyl groups are essential for

the following warm pressing process.

13C MAS NMR (Figure 16(b)) also demonstrates that the pre-crosslinked [-(PhVi)Si-

NH-]n and [-(MeVi)Si-NH-]n still have unsaturated vinyl bonds. The two peaks of =CH- and

=CH2 appear between 130ppm and 140ppm for both the [-(PhVi)Si-NH-]n and [-(MeVi)Si-

NH-]n pre-crosslinked materials. They are located in the same region as in the liquid state

NMR spectra (Figure 15(b)), which were collected on as-synthesized polymers. The residual

unsaturated bonds are important because they help the monolith shaping via warm-presssing.

The peaks at ca. 126 ppm in the 13C solid state MAS NMR spectra of [-(PhVi)Si-NH-]n pre-

crosslinked material are from phenyl groups, and that at 0 ppm are from methyl groups for

both [-(PhVi)Si-NH-]n and [-(MeVi)Si-NH-]n pre-crosslinked materials. The signal is not

intense in the case of [-(PhVi)Si-NH-]n pre-crosslinked material, because the methyl groups

only come from the saturated vinyl groups. The peak is very strong in the spectrum of pre-

crosslinked [-(MeVi)Si-NH-]n since this precursor has a methyl group attached to silicon.

120 100 80 60 40 20 0 -20 -40 -60 -8029Si Chemical shift (ppm)

(a)


250 200 150 100 50 0 -50


Figure 16 (a) 29Si and (b) 13C MAS NMR of pre-crosslinked [-(PhVi)Si-NH-]n (bold line) and [-(MeVi)Si-NH-]n

(thin line). The asterisks in 13C MAS NMR spectra represent spinning side bands

29Si, 13C and 11B MAS NMR (Figure 17) spectra of the boron modified

polyborosilazanes were measured. The peaks at ca. -31 and -19 ppm in the 29Si MAS NMR

spectrum of B[-(C2H4)Si(Ph)-NH-]3n are due to -N-(PhVi)Si-N- cites and -N-Si(PhCsp3)-N-

sites, respectively. Similarly, the peaks at ca. -20 and -7 ppm in the 29Si MAS NMR spectrum

of B[-(C2H4)Si(Me)-NH-]3n are due to -N-(MeVi)Si-N- cites and -N-Si(MeCsp3)-N- sites,

respectively. The signal at ca. -1 ppm for both polymers is due to -N-Si(Csp3)3 cites which

indicate the end groups. The presence of big amount end groups reveals the existence of low

molecular weight borosilazane oligomers in the samples.The 13C MAS NMR spectra of

boron-containing polymers are similar to the pre-crosslinked boron-free polymers. Through

the reaction with borane, vinyl groups are partially saturated. Signals at ~130 and ~140 ppm

are due to the unsaturated carbon double bonds. The saturated –CH2– are manifested as peaks

at ca. 10 ppm. The band near 28ppm is a signal of the saturated ––CH– bonds, and that at

about 0 ppm is from the –CH3bonds. The appearance of saturated vinyl groups confirms the

hydroboration reaction. The most intense bands in the 11B MAS NMR spectra are at ~50 and

0 ppm, which are signals of the C2B-N bonds and C3B←N environments, respectively. BC3

bonds are manifested as bumps at ca. 70 ppm. From the difference in the integrated intensity

beneath these bands, B[-(C2H4)Si(Me)-NH-]3n has more B-C and less B-N bondingthan

B[-(C2H4)Si(Ph)-NH-]3n.

(b)


40 20 0 -20 -4029Si chemical shift (ppm)

200 150 100 50 0 -50


200 150 100 50 0 -50 -100


B[-(C2H4)Si(Ph)-NH-]3n

11B chemical shift (ppm)

Figure 17 (a) 29Si, (b) 13C and (c) 11B MAS NMR of polyborosilazanes

(a)

(b)

(c)


There are two kinds of bonds connecting B and N. The formation of C2B-N bonds suggests dehydrocoupling of B-H and N-H (Figure 18(a)). The appearance of C3B←N environments suggests that boron atoms bond to nitrogen atoms via coordinate covalent bonds (Figure 18(b)). The B-N bonding may impede the formation of branched polymer with big molecular weight. This explains the low melting point of these boron-containing samples.

(a)

(b) Figure 18 (a) dehydrocoupling of B-H and N-H, (b) boron bonds to nitrogen via formation of adducts

Similar results of B-N bonding can also be observed in the FT-IR measurements (Figure 19). The peaks found around 1380 cm-1, 1100 cm-1 and 780 cm-1 are from B-N bonds. B-C bonds are found at ca. 1270 cm-1 and 1020 cm-1. At the same time, N-H bonds, which are located at about 3400cm-1, are also found, meaning N-H groups obviously still exist in the structure. Some of these take part in the reaction with borane. =C-H bonds and C=C bonds, which were found at about 2930 cm-1 and 1600 cm-1, respectively, represented the remaining vinyl groups which did not react with borane.

4000 3500 3000 2500 2000 1500 1000 500

B-C

B[-(C2H

4)Si(Me)-NH-]

3n

B[-(C2H

4)Si(Ph)-NH-]

3n

B-N

Wavenumber (cm-1)

B-CC=CC-H=C-H

B-NN-H

B-N

Figure 19 FT-IR of polyborosilazanes

‐+


As discussed, pre-pyrolysis is necessary for shaping thermoplastic B[-(C2H4)Si(Ph)-

NH-]3n and B[-(C2H4)Si(Me)-NH-]3n. The pre-pyrolyzed materials act as self-fillers in

warm-pressing. Thermalgravimetry was studied from room temperature to 1400⁰C (Figure

20). The mass loss of the polymers due to gas evolution basically occurred in two steps. B[-

(C2H4)Si(Ph)-NH-]3n has a first mass loss step above 200⁰C, while B[-(C2H4)Si(Me)-NH-

]3n begins to lose mass at a much lower temperature. This mass loss is likely due to the

oligomers, which are easily removed during thermolysis. Both B[-(C2H4)Si(Ph)-NH-]3n

and B[-(C2H4)Si(Me)-NH-]3n have a second gas-release step from 400⁰C to 700⁰C. Mass

spectra were measured, as shown in Figure 21, the mass-to-charge ratios (i.e., m/z values) of

these gasses are 78 for C6H6, 52 for C2N2, 41 for CH3-CN, 27 for HCN, 17 for NH3, 16 for

CH4, and 2 for H2. N2 (28) is also detected, but is not given in the figure, because during the

measurements the purge gas (air) was also detected. Thus, the “detected” amount of evolved

N2, which has a mass fragment of m/z=28, was partially from N2 in the purge gas.Self-fillers

were produced by the thermolysis of both polymers at 500C. According to the TG analysis,

at this temperature, a significant amount of gas is released, but the decomposition of the

polymers is not yet complete. Gasses, including H2, CH4 and NH3, keep releasing above 500

C.

200 400 600 800 1000 1200 1400

40

60

80

100

Mas

s (

%)

Temperature (C)

-0,6

-0,5

-0,4

-0,3

-0,2

-0,1

0,0

DT

G (

%/m

in)

Figure 20 TG (solid line) and DTG (dash line) of B[-(C2H4)Si(Ph)-NH-]3n (bold line) and B[-(C2H4)Si(Me)-

NH-]3n (thin line)


200 400 600 800 1000 1200 1400

m/z = 78

m/z = 52m/z = 27

m/z = 17

m/z = 16

m/z = 2

Temperature (C)

Inte

nsi

ty

200 400 600 800 1000 1200 1400

m/z = 17m/z = 41

m/z = 16

m/z = 27

m/z = 2

Temperature (C)

Inte

nsi

ty

Figure 21 Mass loss of (a) B[-(C2H4)Si(Ph)-NH-]3n and (b) B[-(C2H4)Si(Me)-NH-]3n during second gas-

releasing step

MS spectra confirm that the pre-pyrolyzed samples are intermediate materials between

polymer and ceramic—they still maintain a certain amount of organic groups. The remaining

organic groups in pre-pyrolyzed powders, e.g. N-H and C=C-H, offer the opportunity to react

further with as-synthesized polymer. This process helps the achievement of warm-pressing.

FT-IR spectra measured on the 500C pre-pyrolyzed samples are shown in Figure 22. Organic

groups including =C-H, -C-H and N-H can still be detected in the structure. This result is

constant with MS spectra.

4000 3500 3000 2500 2000 1500 1000 500

500C-pyrolyzedB[-(C2H4)Si(Me)-NH-]3n

500C-pyrolyzedB[-(C2H4)Si(Ph)-NH-]3n

C-H=C-H

Wave number (cm-1)

N-H

Figure 22 FT-IR spectra of 500 C pre-pyrolyzed B[-(C2H4)Si(Ph)-NH-]3n and B[-(C2H4)Si(Me)-NH-]3n

Monoliths from boron-free [-(PhVi)Si-NH-]n and [-(MeVi)Si-NH-]n were warm-

pressed using pre-crosslinked materials (as discussed before). Boron-containing B[-

(C2H4)Si(Ph)-NH-]3n and B[-(C2H4)Si(Me)-NH-]3n were warm-pressed using a mixture of

the pre-pyrolyzed powder and the original polymers. The original polymers offer the fluidity

(a) (b)


needed for shaping, while the pre-pyrolyzed powder acted as fillers to form the skeleton of the

compacts. The ratio of the pre-pyrolyzed powder to the polymer was adjusted in order to

obtain a dense greenbody compact without signficant defects. As shown in Figure 23,

essentially defect-free green bodies and pyrolyzed ceramics (at 1100⁰C, Figure 24) were

obtained from all precursors.

Figure 23 warm-pressed greenbodies from the precursors: (a) [-(PhVi)Si-NH-]n, (b) [-(MeVi)Si-NH-]n, (c) B[-

(C2H4)Si(Ph)-NH-]3n and (d) B[-(C2H4)Si(Me)-NH-]3n

Figure 24 (a) SiCN3, (b) SiCN4, (c) SiBCN3 and (d) SiBCN4 bulk ceramics pyrolyzed at 1100⁰C

SEM of the ceramics pyrolyzed at 1100⁰C reveal an essentially defect-free,

homogeneous structure. The corresponding crosssections are shown in Figure 25.

Figure 25 SEM of crosssections of (a) SiCN3, (b) SiCN4, (c) SiBCN3 and (d) SiBCN4 monoliths pyrolyzed at 1100 ⁰C

(d)(c)

(a) (b)

(a) (b) (d)

(a) (b) (c) (d)

(c)


As a short summary, in order to produce ceramic monoliths, the polysilazanes are pre-

crosslinked and warm-pressed. the polyborosilazanes are good binders in plastic forming,

using the pre-pyrolyzed powder as self-fillers. The optimized ratio between original polymer

and pre-pyrolyzed powder was found. Bulk green bodies were formed, and finally, compact

ceramic monoliths were produced after pyrolysis at 1100 ⁰C.

4.1.3 Polysilsesquicarbodiimide and 13C/15N isotope-enriched polysilsesquicarbodiimide

Due to the low natural abundance of 15N (0.37%), it is very difficult to detect the 15N

nucleus by NMR Therefore the nitrogen environments in polymer-derived SiCN ceramics was

previously not well understood. In this work, by enriching the reactant chemicals with the 15N

isotope, it was possible to investigate nitrogen cites in the synthesized polymers and the

derived ceramics.

As shown in Chapter 3, unenriched polyphenylsilsesquicarbodiimide was synthesized

from bis(trimethylsilyl)carbodiimide and trichlorophenylsilane. The 13C/15N isotope-enriched

polyphenylsilsesquicarbodiimide was synthesized from isotope enriched-cyanamide and

trichlorophenylsilane. 13C/15N isotope-enriched polyphenylsilsesquicarbodiimide was

obtained by adding 10% 15N and 13C isotope-labeled cyanamide. The synthesis of the polymer

from the labeled cyanamide is shown in Scheme 15.

Scheme 15 Synthesis of 13C/15N isotope-enriched polyphenylsilsesquicarbodiimide

29Si and 13C MAS NMR (Figure 26) measurements were made on the as-synthesized

polymers. For both the 13C/15N-isotope-enriched and -unenriched [-(Ph)Si-(NCN)1.5-]n

polymer, the 29Si MAS NMR spectra show a main band at -75ppm, which represents the Si

sites in Ph-Si-(NCN)3, confirming the formation of the expected structure. Additionally, [-

(Ph)Si-(NCN)1.5-]n shows two other bands at ca. 0 ppm and -60 ppm. The former band

represents the silicon environments in the starting materials, (CH3)3-Si-NCN-Si-(CH3)3 or ph-

Si-Cl3. The latter band at -60 ppm is assumed to be from the end group Ph-Si(NCN)2-Cl. This

difference in the Si environments between the enriched and unenriched [-(Ph)Si-(NCN)1.5-]n

polymers is due to their different synthesis routes. There is reactant residue of

bis(trimethylsilyl)carbodiimide or trichlorophenylsilane in the unenriched [-(Ph)Si-(NCN)1.5-

]n, while the 13C/15N isotope-enriched [-(Ph)Si-(NCN)1.5-]n from the reaction of cynamide and

trichlorophenylsilane has no such residue. The endgroup Ph-Si(NCN)2-Cl is only observed in

the unenriched polymer. There is no considerable amount of Ph-Si(NCN)2-Cl endgroup in the 13C/15N isotope-enriched polymer. This means that the 13C/15N isotope-enriched polymer may

have a larger average molecular weight.

13C MAS NMR spectra (Figure 26(b)) show that the main bands for both the unlabeled

and labeled polymers lie between 125 and 130 ppm, which represents aromatic carbon rings.


The bands at ca. 120ppm represent carbon in the -N=C=N- environments. This band is much

more enhanced in the 13C/15N isotope-enriched polymer, [-(Ph)Si-(NCN)1.5-]n, because the 13C

isotope is only in the -N=C=N- groups, not in the phenyl groups. The band at ~25ppm in the

isotope-enriched polymer [-(Ph)Si-(NCN)1.5-]n may come from the solvent THF.

15N enrichment of the material offers the opportunity to observe the chemical bonding

environments of nitrogen. Large chemical shifts in nitrogen NMR are attractive from the point

of view of molecular structural investegations. The chemical shifts are very sensitive to

changes in the N environments. 15N MAS NMR of 13C/15N isotope-enriched [-(Ph)Si-

(NCN)1.5-]n is shown in Figure 26 (c). Only one peak at ca. 20ppm is observed in the whole

measuring range. This peak represents N in Si-N=C=N-Si sites.

50 0 -50 -100 -15029Si chemical shift (ppm)

(a)


300 250 200 150 100 50 0 -50


600 400 200 0 -20015N chemical shift (ppm)

Figure 26 (a) 29Si and (b) 13C MAS NMR of [-(Ph)Si-(NCN)1.5-]n(thin line) and 13C/15Nisotope-enriched [-

(Ph)Si-(NCN)1.5-]n (bold line); (c) 15N MAS NMR of 13C/15N isotope-enriched [-(Ph)Si-(NCN)1.5-]n. The

asterisks in 13C MAS NMR spectra represent spinning side bands

13C/15N isotope-enriched [-(Ph)Si-(NCN)1.5-]n polymer-derived ceramics pyrolyzed at

800 and 1100 ºC have one main band at ca. -47ppm (Figure 27), which represents the Si-N4

tetrahedral environment of Si. The lack of any SiCxN4-x (x=0-4) mixed units is characteristic

of polysilylcarbodiimide-derived ceramics.

(c)

(b)


100 50 0 -50 -100 -150 -200

Iso_SiCN5-800 Iso_SiCN5-1100


Figure 27 29Si MAS NMR of Iso_SiCN5 ceramics prepared at 800 ºC(grey line) and 1100 ºC(black line)

The 13C MAS NMR spectra (Figure 28) show a significant difference between the

ceramics derived from the labeled and unlabeled polymers. The carbon atoms in SiCN5

ceramics basically assume only one type of environment, which is graphitic carbon with a

signal centered at ~ 110 ppm. Iso_SiCN5 ceramics show two additional bands at ~125 ppm

and ~150 ppm in the 13C MAS NMR spectra. The band at ~150 ppm represents the sp2C-N

bonding and the one at ~125 ppm is from amorphous sp2 C. It is belived that these two kinds

of carbon environments also exsist in the SiCN5 ceramics which are derived from unenriched

polymer. They do not appear in the spectra because of their relatively low intensity compared

to that of graphitic carbon. However, in the spectra of Iso_SiCN5 ceramics, which are

enriched with 13C, the intensity of the sp2C-N and sp2C bands are relatively enhanced. This

means that the graphitic carbon in the ceramics is mainly derived from the phenyl groups,

while carbon in the sp2C-N and amorphous sp2C environments are mainly derived from

N=C=N groups in the preceramic polymer.


300 250 200 150 100 50 0 -50 -100

SiCN5

sp2C- N


graphitic carbon

amorphous sp2C

Iso_SiCN5

Figure 28 29Si MAS NMR of SiCN5 and Iso_SiCN5 ceramics prepared at 800 ºC (grey line) and 1100 ºC (black

line)

For the isotope-enriched ceramic prepared at 800 ºC, there are two bands in the 15N

MAS NMR spectrum (Figure 29) at 21ppm and 49 ppm, representing N-Si2C and N-Si3,

respectively. When heated to 1100 ºC, a dominant band appears at 49 ppm, which suggests

the formation of N-Si3. For both pyrolysis temperatures (800 and 1100 ºC), 15N CPMAS

NMR shows an intense band at 21ppm, which represents N-Si2C. Thus, a greater amount of

hydrogen atoms are located in the N-Si2C environment. More interestingly, an energetic

investigation reveals that the presence of hydrogen near the N mixed-bond environment

stabilizes the structure of this branched polymer-derived ceramic. Details will be discussed in

Section 4.3.


200 150 100 50 0 -50 -100 -150 -200

N-Si2C

15N chemical shift (ppm)

N-Si3

200 150 100 50 0 -50 -100 -150 -200

N-Si3N-Si2C

15N chemical shift (ppm)

Figure 29 15N (a) MAS and (b) CPMAS NMR spectra of Iso_SiCN5-800(grey line) and Iso_SiCN5-1100(black

line) ceramics

(a)

(b)


4.1.4 Summary

Two classes of preceramic polymers have been synthesized: polysilylcarbodiimide and

polysilazane. Polymers with phenyl groups in the structure had higher carbon contents, while

those with methyl group had lower carbon contents. The formation of these polymers was

confirmed by NMR spectroscopy. Borane was reacted with all obtained polymers to produce

boron-modified counterparts, namely polyborosilylcarbodiimide and polyborosilazane. The

pre-crosslinking process is considered to be the most critical step in shaping the

polysilylcarbodiimides and polysilazanes to manufacture bulk ceramics. Meanwhile, for

polyborosilazanes, it is important to have a pre-pyrolysis procedure. The pre-pyrolyzed

material acts as self-filler, and the original thermoplastic polymers acts as binders. Final bulk

ceramic compacts were successfully produced from all of the precursors. In addition to the

linear polymers, branched polysilsesquicarbodiimide was synthesized, as well as its 13C and 15N isotope-enriched analogue. 15N enrichment of the polymer enables the measurement of 15N MAS NMR and 15N CPMAS NMR spectra which provides more detailed information

related to structural features.

52 Effect of processing route

4.2 Effect of processing route

The polymer-derived ceramic technique is an efficient and promising method to

produce complex-shaped bulk ceramics by taking advantage of plastic technologies [1,39,40].

Uniaxial warm pressing is one of the most frequently used processes in the fabrication of bulk

ceramics from precursors. In this work, we investigated the effect of processing route, namely

to produce powders or bulk ceramics, on the nanostructure of carbon-rich SiCN and SiBCN

polymer-derived ceramics. These ceramics were also characterized by elemental analysis,

SEM, and local analysis by high resolution TEM (HRTEM). The final microstructures are

discussed with respect to the precursor chemistry, processing route (powder and bulk

synthesis), temperature, and presence of boron and “free” carbon.

4.2.1 Effect of processing route on the nanostructure — Low temperature thermal

transformation

The synthesis and processing of the polyphenylvinylsilylcarbodiimide and

polyborophenylsilylcarbodiimide were carried out as described in the Experimental Section,

and are summarized in Figure 30. As discussed in Section 4.1, a pre-crosslinking step is

essential for [-(PhVi)Si-NCN-]n, and after that, green body compacts were produced.

Figure 30 Synthesis and processing of polyphenylvinylsilylcarbodiimide and polyborophenylsilylcarbodiimide. Reprinted from Journal of the European Ceramic Society, V 32, Y. Gao, G. Mera, H. Nguyen, K. Morita, H.-J. Kleebe, and R. Riedel, Processing route dramatically influencing the nanostructure of carbon-rich SiCN and SiBCN polymer-derived ceramics Part I: Low temperature thermal transformation, 1857, copyright (2012), with permission from Elsevier [134].

Two pyrolysis temperatures were selected. 1100 °C, the standard temperature for

amorphous PDC fabrication, was selected since it is the temperature at which the polymer-to-

ceramic transformation has basically completed. 1400°C was selected as the temperature at

which the PDCs just begin to undergo nano-crystallization and show a difference from the

amorphous architecture.

Effect of processing route 53

The microstructure of the green bodies and of the pyrolyzed bulk ceramics were observed and

studied by SEM. A SiCN(O) surface layer, which contains cracks and pores, was found on the surface

of bulk SiCN1-1100 and SiBCN1-1100 ceramics. Figure 31 shows the presence of this ~3 μm layer on

SiCN1-1100. Beneath the SiCN(O) layer, the ceramic is very dense, showing no defects. The same

microstructural characteristics were found also in the SiBCN1-1100 sample. The formation of the

SiCN(O) layer is due to the handling of green body in air. Samples were polished prior to annealing at

1400 °C in order to avoid the influence of the SiCN(O) layer on the microstructure of the bulk

ceramics.

Figure 31. SiCN1-1100 cross-section: a) and b) thin SiCN(O) layer on the surface; c) defect-free bulk SiCN1-1100 ceramic

under the SiCN(O) layer. Reprinted from Journal of the European Ceramic Society, V 32, Y. Gao, G. Mera, H.

Nguyen, K. Morita, H.-J. Kleebe, and R. Riedel, Processing route dramatically influencing the nanostructure of

carbon-rich SiCN and SiBCN polymer-derived ceramics Part I: Low temperature thermal transformation, 1857,

copyright (2012), with permission from Elsevier [134].

FTIR is an important integral method for the characterization of bonding, and

furthermore, the microstructure of the ceramics. Figure 32 presents the ARO FTIR (all-

reflecting objective FTIR) of [-(PhVi)Si-NCN-]n- and B[-(C2H4)Si(Ph)-NCN-]3n-derived

SiCN1 and SiBCN1 bulk and powder samples.

As shown by FTIR, the majority of the bands correspond to the “free” carbon phase

(1652, 1583, 1503, 1447, 1378 cm-1). All ceramics still contain C(sp3)-H bands (C-H

asymmetric and symmetric stretching vibrations at 2924 and 2860 cm-1, respectively)

probably as terminally saturated groups in the graphene-like carbon phase. The most

important bands are related to the presence of the Si3N4 and SiC phases (Si-N stretching

vibration at 970 cm-1, and Si-C stretching at 812 cm-1). Remarkably, the FTIR of all samples

contain C=N (stretching vibration at 1720 cm-1) and C-N (stretching at 1253 and 1188 cm-1).

Moreover, the FTIR spectrum of SiCN1-1400 powder shows an intense band for Si-C due to

the crystallization of SiC (812 cm-1). The intensity of the Si3N4 band also varies depending on

the temperature and processing route. For the bulk ceramics, the silicon nitride band is strong,

independent of the annealing temperature. In the case of the powder ceramics, at 1400 °C, the

Si-N band is less intense than in the 1100 °C analogue. This observation corroborates very

well with the increase in the intensity of the SiC band. The ceramics are losing nitrogen by the

carbothermal reaction of Si3N4 with carbon to form SiC and nitrogen gas according to a-Si3N4

+ 3 a-C 3 α/ß-SiC + 2 N2 .


Figure 32 ARO FTIR spectroscopy of (a) SiCN1 ceramics and (b) SiBCN1 ceramics at 1100 and 1400 °C. Reprinted from

Journal of the European Ceramic Society, V 32, Y. Gao, G. Mera, H. Nguyen, K. Morita, H.-J. Kleebe, and R.

Riedel, Processing route dramatically influencing the nanostructure of carbon-rich SiCN and SiBCN polymer-

derived ceramics Part I: Low temperature thermal transformation, 1857, copyright (2012), with permission from

Elsevier [134].

Raman spectroscopy is a key nondestructive tool for the examination of the structural

evolution of the “free” carbon phase in PDCs. [138–142] The representative features of

“free” carbon in the Raman spectra of PDCs are (1) the so-called disorder-induced D-band at

approx. 1350 cm-1, and (2) the G band at approximately 1582 cm-1 due to in-plane bond

stretching of sp2 carbon, and (3) the G´-band (the overtone of the D-band which is always

observed in defect-free samples) at 2700 cm-1. The D and G bands can vary in intensity,

position, and width, depending on the structural organization of the sample under

investigation. The intensity ratio of the D and G modes, ID/IG, enables the evaluation of the

lateral size of graphitic carbon using the formula (Formula 3) reported by Ferrari and

Robertson: [138]

2)(' aG

D LCI

I

Formula 3

where, La is the size of graphitic carbon along the six fold ring plane (lateral size), and C´ is a

coefficient that depends on the excitation wavelength of the laser. The value of the

coefficient, C´, for the wavelength of the Ar-ion laser employed (514.5 nm) is 0.0055 Å-2.

Gaussian curve fitting of the Raman bands was performed in order to extract the ID/IG

intensity ratios, and to determine the lateral size of the graphitic carbon. Peak fitting was

completed including the minor T bands (shoulder at ~1200 cm-1), which are attributed to sp2-

sp3 C-C and C=C bonds, and the D´´ at ~1500 cm-1 corresponding to the fraction of

amorphous carbon contained in the samples [143]. These additional bands were present due to

the structural disorder that activates otherwise forbidden vibrational modes [144,145].

(a) (b)


Figure 33 Micro-Raman spectroscopy of a) SiCN ceramics (a1) SiCN1-1100 powder; (a2) SiCN1-1400 powder; (a3) SiCN1-

1100 bulk; (a4) SiCN1-1400 bulk; and b) SiBCN ceramics: (b1) SiBCN1-1100 powder; (b2) SiBCN1-1400 powder; (b3)

SiBCN1-1100 bulk; (b4) SiBCN1-1400 bulk. The corresponding cluster size of carbon (La) for each measurement is provided

in the graphs. Reprinted from Journal of the European Ceramic Society, V 32, Y. Gao, G. Mera, H. Nguyen, K.

Morita, H.-J. Kleebe, and R. Riedel, Processing route dramatically influencing the nanostructure of carbon-rich

SiCN and SiBCN polymer-derived ceramics Part I: Low temperature thermal transformation, 1857, copyright

(2012), with permission from Elsevier [134].

The micro-Raman spectra of all samples are characterized by broad and overlapped D

and G bands, indicating a strongly disordered state (Figure 33). The disorder in the carbon

phase can primarily be induced by the presence of edges in the graphene layers, by the

deviation from planarity of the graphene layers, and also by the presence of carbon atoms in

the sp3 hybridization state. Recently, the dependence of the ID/IG ratio on the degree of

disorder in graphene-like materials was reported [146–148]. The disorder was quantified as

depending on point-like defects. Assuming that we have a low defect density, the typical

interdefect distance, LD, can be calculated with the formula: ID/IG = C(λ)/LD2, which is valid

when LD > 6 nm (C(514.5 nm) ≈ 107 nm2) [146–148]. Unfortunately, the equation is not

valid for high defect densities, but despite this, it can help qualitatively compare the samples

prepared by different processing routes. The processing of bulk ceramics is expected to result

in a decreased defect densitiy in the carbon phase. The interdefect distances, LD, together with

the ID/IG ratio and the lateral cluster size La are summarized in Table 4. Except for the

SiBCN1-1100 powder sample, all other ceramics show valid LD distances. For the SiBCN1-

1100 powder sample, the distance between defects (LD = 5.71 nm) is lower than the limit of

validity of the equation (LD > 6 nm), and is therefore omitted from discussion. As observed in

Table 1, the bulk samples show a lower defect density than the powder samples. Moreover, at

1400 °C, the carbon phase is more organized than at 1100 °C, and also shows fewer defects.

Furthermore, the presence of boron decreases the defect density and enhances the

organization of the carbon. Especially at 1100 °C, the SiBCN1-1100 bulk sample shows

decreases in the lateral size of the carbon (1.41 nm) and in the ID/IG ratio due to the high

organization of the graphene layers. This fact is underlined also by the presence of a sharp

single Gaussian peak for the G´ band at 2672.62 cm-1. Regarding the lateral size of the

graphitic carbon in the SiCN ceramics, a decrease is observed when comparing the 1400 °C

(a) (b)


samples to the 1100 °C ones. This trend is observed in both the powder and bulk ceramics

cases.

Table 4 The Raman parameters of the “free” carbon phase in powder and bulk SiCN1 and SiBCN1 ceramics. Sample ID/IG La [nm] LD [nm]

SiCN1-1100 powder 2.61 2.18 6.40 SiCN1-1400 powder 1.71 1.76 7.91 SiCN1-1100 bulk 2.54 2.15 6.49 SiCN1-1400 bulk 1.92 1.87 7.47 SiBCN1-1100 powder 3.28 2.44 5.71 SiBCN1-1400 powder 2.43 2.10 6.64 SiBCN1-1100 bulk 1.09 1.41 9.90 SiBCN1-1400 bulk 2.61 2.18 6.40

The compositions of the SiCN1 and SiBCN1 bulk and powder samples were

determined by elemental analysis (Table 5).

Table 5 Chemical composition of the powder and bulk SiCN1 and SiBCN1 ceramics.

Sample Chemical composition

Empirical formula Si % C % N % B % O % H % Cl %

SiCN1-1100 Powder

24.80 56.50 14.80 - 2.14 0.24 1.55 SiC5.32N1.19H0.27Cl0.05O0.15

SiCN1-1100 Bulk

28.45 51.42 10.98 - 9.15 - - SiC4.22N0.77O0.56

SiBCN1-1100 Powder

23.79 51.32 16.58 3.05 2.21 0.30 2.75 SiB0.33C5.04N1.40H0.35Cl0.09O0.16

SiBCN1-1100 Bulk

28.81 42.31 13.63 3.41 11.84 - - Si1B0.30C3.43N0.88O0.72

SiCN1-1400 Powder

29,24 59.21 11.03 - 0.52 - - SiC5.47N0.87O0.04

SiCN1-1400 Bulk

28.65 57.06 11.66 - 2.63 - - SiC4.65N0.81O0.16

SiBCN1-1400 Powder

28.21 51.08 16.19 3.06 1.46 - - SiB0.28C4.23N1.07O0.09

SiBCN1-1400 Bulk

29.15 48.62 14.62 3.45 4.16 - - SiB0.30C3.89N0.94O0.25

Regarding the compositions of the powder samples compared to their bulk analogues,

it was observed that the powder ceramics at 1100 °C and 1400 °C contain a higher content of

carbon and nitrogen. At the same time, the silicon content is higher in the bulk ceramics.

Another interesting observation is that the hydrogen and chlorine contents registered in the

powder ceramics at 1100 °C are not present in the bulk ceramics pyrolyzed at the same

temperature. Even though the FTIR study showed the presence of C-H bonds in bulk

ceramics, the amount of hydrogen is less than the detection limit for elemental analysis.

Differences in the Si, C, N, H and Cl contents are attributed to changes in the chemistry and

architecture of precursors using different processing routes. The processing of bulk ceramics

includes a crosslinking and shaping step, which allows the formation of a 3D structure, and

the elimination of the chlorine-containing end-group from the precursor. The boron content in

all SiBCN samples is constant, with no difference measured between the bulk and powder


ceramics. As presented in Figure 31, the bulk SiCN1 and SiBCN1 samples pyrolyzed at 1100

°C show a 3 m thick SiCN(O) surface layer. In order to understand how the presence of this

layer can change the composition of the samples, elemental analysis was completed without

polishing. As expected, the SiCN1 1100 °C bulk and SiBCN1-1100 bulk contain a high

amount of oxygen contamination—9.15 and 11.84 %, respectively. The oxygen-containing

surface layer was removed from the 1400 ºC bulk analogues prior to elemental analysis, and

thus these samples showed less oxygen contamination.

High-resolution TEM microscopy (HRTEM) is a powerful method for locally

investigating the nanostructural evolution of polymer-derived ceramics at different

temperatures of annealing. Even if the PDCs are X-ray amorphous after pyrolysis at low

temperatures (T < 1400 °C), they are typically heterogeneous as shown by several TEM

studies [149–151]. Until now, there existed a lack of information regarding the TEM

investigations of polysilylcarbodiimides-derived ceramics. Due to the complicated, and until

now, never-before successful processing of these precursors into bulk monoliths, no

nanostructure investigation of these ceramics has been reported. This study presents an easy

route to produce bulk carbon-rich polysilylcarbodiimides-derived ceramics, and their

nanostructural analysis by TEM. By comparing bulk and powder SiCN ceramics, the effects

of the crosslinking and shaping processes on the thermal stability of PDCs can be indirectly

assessed. Moreover, the low boron content of these samples is discussed in light of the

resultant nanostructure.

Figure 34 presents the HRTEM images of the powder and bulk SiCN and SiBCN

ceramics at 1100 °C. The selected area electron diffraction (SAED) patterns of the SiCN and

SiBCN ceramics synthesized at 1100 °C are displayed as insets in the figures. All four

ceramics, independent of composition or processing route, show a diffuse, elastically

scattered ring pattern typical of amorphous samples. No major contrast variations were

observed for the samples annealed at 1100 °C, indicating the amorphous nature of the

samples. At even higher magnifications, no indication of any crystalline phases could be

imaged. At this low pyrolysis temperature, no difference can be observed between the

nanostructures of the powder and bulk SiCN ceramics, with or without boron.


Figure 34. HRTEM images of powder and bulk SiCN and SiBCN ceramics at 1100 °C (the insets present the selected area

electron diffraction): (a) SiCN1-1100 powder; (b) SiCN1-1100 bulk; (c) SiBCN1-1100 powder; (d) SiBCN1-1100 bulk.

Reprinted from Journal of the European Ceramic Society, V 32, Y. Gao, G. Mera, H. Nguyen, K. Morita, H.-J.

Kleebe, and R. Riedel, Processing route dramatically influencing the nanostructure of carbon-rich SiCN and

SiBCN polymer-derived ceramics Part I: Low temperature thermal transformation, 1857, copyright (2012), with

permission from Elsevier [134].

In the case of powder SiCN1 and SiBCN1 ceramics pyrolyzed at 1400 °C, high-

resolution TEM imaging shows that crystallization has initiated in the samples. Figure 35

displays the nanostructural characteristics of these ceramics, in which crystallites of α/ß-SiC

are embedded in a graphene-like carbon matrix. In addition, as indicated by the elemental

analysis, amorphous Si3N4 is still present in the amorphous network (Table 5). As previously

reported for carbon-rich SiCN ceramics obtained from polysilylcarbodiimides, no Si3N4

crystallization was detectable [98]. As observed in the SAED insets in Figure 35, the SiBCN1

sample (Figure 35b) shows enhanced crystallization compared to the SiCN1 (Figure 35a)

analogue. Thermodynamic modeling studies as well as experimental work demonstrate that

the presence of boron in SiCN is the driving force for SiC crystallization [15,152]. Indeed,


also in the case of the polyborophenylvinylsilylcarbodiimide-derived SiBCN1 ceramic, the

same enhancement in SiC crystallization due to the presence of boron was observed.

Figure 35. HRTEM images of powder SiCN and SiBCN ceramics annealed at 1400 °C. The SAED patterns (insets) show the

presence of α/ß-SiC. Reprinted from Journal of the European Ceramic Society, V 32, Y. Gao, G. Mera, H. Nguyen,

K. Morita, H.-J. Kleebe, and R. Riedel, Processing route dramatically influencing the nanostructure of carbon-

rich SiCN and SiBCN polymer-derived ceramics Part I: Low temperature thermal transformation, 1857,

copyright (2012), with permission from Elsevier [134].

In contrast to the nanostructures of the SiCN and SiBCN powder ceramics, the TEM

investigation of the bulk analogues revealed no α/ß-SiC crystallization. Even at higher

magnifications, no SiC nuclei were detectable. In order to confirm the amorphous nature of

the bulk samples, several TEM micrographs at different defocus settings of the objective lens

were taken.


Figure 36 HRTEM images of bulk SiCN and SiBCN ceramics pyrolyzed at 1400 °C. Different defocus setting of the

objective lens were chosen since carbon and SiC can better be imaged at slight over- and underfocus conditions [153],

respectively; (a), (b), (c) SiCN 1400 °C bulk and (d), (e), (f) SiBCN 1400 °C bulk. The insets represent the inverse FFT

images of the selected areas (boxed regions). Reprinted from Journal of the European Ceramic Society, V 32, Y. Gao,

G. Mera, H. Nguyen, K. Morita, H.-J. Kleebe, and R. Riedel, Processing route dramatically influencing the

nanostructure of carbon-rich SiCN and SiBCN polymer-derived ceramics Part I: Low temperature thermal

transformation, 1857, copyright (2012), with permission from Elsevier [134].

For polysilazanes-derived ceramics, with increasing temperature, the first

crystallization event is the formation of basic structural units (BSU) of carbon, followed by

the nucleation of SiC. In the SiCN bulk ceramic annealed at 1400 °C, depending on the

defocus setting of the objective lens, slight contrast variations are seen in the HRTEM image.

At an under- (-20 nm) and an overfocus (+20nm) setting (Figure 36), the inverse fast Fourier

transform (FFT) images show variations in phase contrast, indicating a fine organization of

the carbon phase. The FFT inset clearly shows the presence of a carbon phase (Figure 37a), as

also indicated in the inverse FFT image of a thicker region (Figure 37b).


Figure 37 (a) HRTEM micrographs of SiCN1-1400 bulk (the inset shows the FFT image corresponding to carbon phase); (b)

inverse FFT image of the thick region of SiCN1-1400 bulk. Reprinted from Journal of the European Ceramic Society,

V 32, Y. Gao, G. Mera, H. Nguyen, K. Morita, H.-J. Kleebe, and R. Riedel, Processing route dramatically

influencing the nanostructure of carbon-rich SiCN and SiBCN polymer-derived ceramics Part I: Low

temperature thermal transformation, 1857, copyright (2012), with permission from Elsevier [134].

In the case of bulk SiBCN pyrolyzed at 1400 °C (Figure 36d-f), depending on the

defocus value of the objective lens, no phase contrast variations can be seen in the HRTEM

images, with the sample being fully amorphous. This fact can be due to the higher thermal

stability of the SiBCN ceramics. It has been assumed that the boron-containing carbon matrix,

similar to a graphene-like phase, acts as an effective diffusion barrier in preventing SiC

nucleation. As presented in Table 5, the SiBCN ceramics, both powder and bulk, contain only

up to 3-3.4 wt% of boron—an amount obviously not sufficient to prevent SiC crystallization

in higher surface area powder samples.

Differences in thermal stability against crystallization between the powder and bulk

samples are due to surface nucleation and increased nanoporosity in the powders. The greater

specific surface area of powder samples increases the “reactivity” of the materials with regard

to the carbothermal reaction. Thus, the powders begin to crystallize at lower temperatures

and at higher rates. A confirming trend was previously observed for SiBCN ceramics of

different particle sizes as investigated with respect to thermal decomposition and

crystallization [33]. The skin-core effect reported for polyborosilazanes-derived

ceramics [34] was not observed in our study.

It has been reported that, in the case of SiBCN ceramics composed of Si3N4, SiC, BN

and C, thermal stability strongly depends on the boron content [154,155]. In our case, a low

boron concentration is expected to decrease the thermal stability of the ceramics against

crystallization. However, at the same time, a high “free” carbon concentration is assumed to

improve the thermal stability of these materials, counterbalancing the boron effect [17–19].

Furthermore, surfaces provide active reaction sites for decomposition and crystallization.


4.2.2 Effect of processing route on the nanostructure — Thermal transformation up to

2100°C

Thermal transformation of SiCN1 and SiBCN1 ceramics (powder and bulk) are further

studied at high temperatures. Figure 38 compares the TG curves of powder and bulk SiCN1-

1100 and SiBCN1-1100 samples as heated up to 2100C. Several interesting features can be

seen from the curves. For the SiCN1 material, the weight loss of the powder sample is

drastically different from that of the bulk sample. At up to 2100oC, the bulk sample only loses

about 16.0 % of its original weight, while the powder sample loses 26.1 wt% (Figure 38 a).

On the other hand, the SiBCN1 powder and bulk samples exhibited similar amounts of total

weight loss at 2100oC (~ 26.5 % of their original weight) (Figure 38 b), even though the bulk

sample loses less weight than the powder one at temperatures below 1900oC. The SiCN1 and

SiBCN1 powder samples exhibited similar weight loss behavior: both of them lose weight

significantly between 1400 and 1650oC, and almost no weight change occurs at higher

temperatures. The total amount of weight loss at 2100oC is also similar for these samples.

However, the weight loss of bulk SiCN1 is much less than that of bulk SiBCN1. Closer

comparison between the two curves reveals that both SiCN1 and SiBCN1 samples exhibit

similar weight loss below 1400oC. At higher temperatures, both samples exhibit two weight

loss stages at 1400-1600oC, and at 1600-1900oC, respectively. During both stages

(particularly, during the second stage), SiBCN1 exhibited more weight loss than SiCN1.

These results clearly show that bulk samples have improved thermal stability against weight

loss, which is more obvious for SiCN1 than for the SiBCN1, and that boron likely decreases

the thermal stability of bulk samples. However, boron has no noticeable effect on the weight

loss of powder samples.

1000 1200 1400 1600 1800 200070

75

80

85

90

95

100

Mas

s (

%)

Temperature (C)

SiCN1-1100 Bulk SiCN1-1100 Powder

1000 1200 1400 1600 1800 200070

75

80

85

90

95

100

Mas

s (%

)

Temperature (C)

SiBCN-1100 Bulk SiBCN-1100 Powder

Figure 38: TG cures of a) SiCN1-1100 and b) SiBCN1-1100 bulk (solid line) and powder (dash line) samples.

The thermal transform behavior of the SiCN1 and SiBCN1 ceramics were further

analyzed by comparing their differential TG (DTG) curves (Figure 39). It is seen that the

DTG curves for the SiCN1 and SiBCN1 powder samples exhibit one peak centered at ~

1540oC, while the DTG curves for the bulk SiCN1 and SiBCN1 ceramics showed two peaks

centered at ~ 1540 and 1750oC, respectively. For the bulk samples, the first peak occurs at the

(a) (b)


same temperature as the single peak for the powder samples, indicating they arise from the

same reaction. The additional peak of the bulk samples occurs at higher temperature,

suggesting the occurance of an additional degradation reaction in the bulk samples.

1000 1200 1400 1600 1800 200070

75

80

85

90

95

100

TG

(%

)

Temperature (C)

-0,0012

-0,0010

-0,0008

-0,0006

-0,0004

-0,0002

0,0000

0,0002

0,0004

0,0006

DT

G

TG

DTG

1000 1200 1400 1600 1800 200080

85

90

95

100

TG

(%

)

Temperature (C)

0,000

DTG

DT

G

TG

1000 1200 1400 1600 1800 200070

75

80

85

90

95

100

DTG

TG

(%

)

Temperature (C)

TG

-0,0012

-0,0010

-0,0008

-0,0006

-0,0004

-0,0002

0,0000

0,0002

0,0004

DT

G

1000 1200 1400 1600 1800 200070

75

80

85

90

95

100T

G (

%)

Temperature (C)

-0,0012

-0,0010

-0,0008

-0,0006

-0,0004

-0,0002

0,0000

0,0002

0,0004

0,0006

DTG

DT

G

TG

Figure 39: HT-TG and DTG curves of a) SiCN1-1100 powder, b) SiCN1-1100 bulk, c) SiBCN1-1100 powder and d)

SiBCN1-1100 bulk samples.

Two annealing temperatures were selected: 1700oC, which is a temperature between

the two peaks of the DTG curves of the bulk samples, and 2000oC, which is a temperature

above both peaks. The compositions of the SiCN and SiBCN ceramics were measured by

elemental analysis on both powder and bulk samples annealed at various temperatures (Table

6). It is interesting to see that, for the SiCN materials, the powder samples always contained

higher carbon and silicon contents, while the nitrogen and oxygen contents are lower than

bulk samples annealed at the same temperature. In contrast, for the SiBCN materials, the

powder samples always contained higher carbon contents, while the silicon, nitrogen and

oxygen contents were lower than in the bulk samples. Boron content remains the same for all

samples, regardless of form and annealing temperature.

(a) (b)

(c) (d)


Table 6: Chemical composition of the powder and bulk SiCN1 and SiBCN1 ceramics pyrolyzed at 1700 and 2000 oC

Sample Chemical composition (at. %)

Empirical formula Calculated phases (Mol%)

Si C O N B Free C Si3N4 SiC BN SiO2

SiCN1-1700 powder

33.31 65.38 0.71 0.60 / SiC4.59N0.04O0.04 78.69 0.20 20.71 / 0.40

SiCN1-2000 powder

32.25 67.75 0 0 / SiC4.91 79.60 0 20.40 / 0

SiCN1-1700 bulk

26.14 62.19 1.68 9.99 / SiC5.56N0.76O0.11 89.35 3.30 6.39 / 0.97

SiCN1-2000 bulk

31.12 63.18 0.42 5.29 / SiC4.75N0.34O0.02 82.83 1.76 15.17 / 0.24

SiBCN1-1700 powder

29.53 61.54 1.03 7.91 3.36 SiB0.30C4.87N0.54O0.06 77.62 1.15 15.03 5.62 0.58

SiBCN1-2000 powder

31.08 62.03 0 6.89 3.83 SiB0.32C4.67N0.44 74.89 0.62 18.12 6.38 0

SiBCN1-1700 bulk

31.81 54.33 2.91 10.95 3.54 SiB0.30C3.99N0.69O0.16 75.56 2.24 13.92 6.48 1.80

SiBCN1-2000 bulk

31.17 58.64 1.27 8.92 3.83 SiB0.32C4.40N0.57O0.07 75.21 1.32 16.10 6.63 0.74

It is well-accepted that the transformation of polymer-derived amorphous ceramics at

high annealing temperatures involves several reactions, including:

Si3+xN4Cx+y Si3N4 + xSiC + yC (1)

2Si-N + 2C 2Si-C + N2 (2)

2Si-N 2Si + N2 (3)

Si + C SiC (4)

The first reaction starts at low temperatures, and leads to phase seperation. The

carbothermal reaction (Reaction (2)) and decomposition of Si3N4 (Reaction (3)) occur at 1484

and 1841oC under 1 atm N2 atmosphere, respectively. The product Si from Reaction (3) may

react with carbon again and produce SiC(Reaction (4)).

Previous studies revealed that the carbothermal reaction (Reaction (2)) and the

following crystallization of SiC in polymer-derived amorphous ceramics preferentially

initiates at surfaces. We tested this theory by annealing polished bulk samples at 1400oC.

Figure 40 compares the XRD patterns of the SiCN1 and SiBCN1 ceramics before and after

surface polishing. SiCN1 and SiBCN1 bulk ceramics without surface polishing are marked as

SiCN1_S and SiBCN1_S. It is clearly seen that, for both SiCN1 and SiBCN1 ceramics,

crystallization iniates at 1400oC, and is only observed at the surface.


10 20 30 40 50 60 70 80 90

SiBCN1-1400

SiCN1-1400

SiBCN1-1400_S

In

tens

ity (

a.u.

)

2

SiC

SiCN1-1400_S

Figure 40: XRD patterns of bulk ceramics prepared at 1400C before and after polishing

It is seen that Reaction (1) does not lead to any weight change; while both Reactions

(2) and (3) result in weight loss. Thereby, we can conclude that the weight loss exhibited by

the current materials is mainly due to the evolution of N2. Indeed, the nitrogen content within

the materials significantly decreased with increasing annealing temperature, which can be

seen more clearly by comparing the nitrogen content in the samples annealed at 1700oC with

their analogues annealed at 1400oC, as discussed in Section 4.2.1 [134]. Since the TG

analyses in this study were carried out in a pure argon atmosphere, the reaction (2) and (3) are

encouraged and their respective starting temperatures may shift to lower temperatures.

Starting temperatures of Reaction (2) and Reaction (3) shift to below 1400 oC and 1800 oC,

respectively. For the powder samples, due to the high surface-to-volume ratio, reaction (2)

can quickly occur on the surface of the powders at relative low temperature (~1400-1700 oC),

leading to a single peak in the corresponding DTG curves (Figure 39 a and c). The boron

effect is not obvious. For the bulk samples, Reaction (2) occurs on the surface. Si3N4 in the

inner part of bulks remains unreacted. Plateaus are shown in temperature range ~ 1600-1700

oC in the TG curves of bulks. This means that there is no significant N2 revolution in this

temperature range. With increasing temperature, Reaction (3) occurs and the bulks lose

weight again. Consequently, they exhibit two weight loss peaks (Figure 39 b and d). The

build-up of N2 pressure within the sample may impede the decomposition of Si3N4 (Reaction

(3)). This explains the lesser weight loss of the bulk samples as compared to the powder

samples, especially in the case of SiCN ceramics.

While there is no considerable difference in weight loss between the SiCN1 and

SiBCN1 powder samples, the SiCN1 bulk sample showed dramatically greater thermal

stability against weight loss as compared to the SiBCN1 analogue. Closer comparison of the

TG curves reveals that the weight-loss difference between the SiCN1 and SiBCN1 mainly


occurs at the higher temperature range (1600-2100oC), suggesting that Reaction (3) is the

dominate source. While Reaction (3) occurres in both bulks, it undergoes faster in SiBCN1

bulk than in SiCN1 bulk. The reason for this is not clear at this moment. One possible

explanation is that the boron-containing SiBCN1 bulk sample has more pores than the SiCN1.

Therefore, N2 is released more easily in SiBCN1 bulk and Reaction (3) is enhanced.

The thermal stability of the materials against crystallization was first analyzed using

XRD. The results (Figure 41a) reveal that, in SiCN1, the powder sample annealed at 1700oC

contains SiC as the main crystalline phase, together with a very small amount of Si3N4. For

the 2000oC annealed powder sample, SiC is still the dominant crystalline phase, in addition,

there is also a small amount of crystalline graphite which is not found in the 1700oC annealed

sample. On the other hand, the SiCN1 bulk samples contain more Si3N4 phase than the

powder sample; and the concentration of the Si3N4 phase in the bulk sample increases with

increasing annealing temperature. No crystalline carbon was found in bulk samples.

10 20 30 40 50 60 70 80 90

SiCN1-2000 bulk

SiCN1-1700 bulk

SiCN1-2000 powder

••

•

••

•

•

• ••

•

Si3N

4

Si3N

4

• SiCgraphite

Inte

nsi

ty (

a.u

.)

2

• ••

• SiCN1-1700 powder

(a)


10 20 30 40 50 60 70 80 90

SiBCN1-2000 bulk

SiBCN1-1700 bulk

SiBCN1-2000 powder

Si3N

4

Si3N

4

SiC graphite

Inte

nsi

ty (

a.u

.)

2

SiBCN1-1700 powder

Figure 41: XRD patterns of (a) SiCN1 and (b) SiBCN1 ceramic powder and bulk samples at 1700C and 2000C

Figure 41(b) shows the XRD patterns for the SiBCN1 samples. Similar to SiCN1,

SiBCN1 powder samples also have a major SiC phase occurring with a very small amount of

Si3N4. The amount of the Si3N4 phase increases with increasing annealing temperature, and a

small amount of graphite is found in the 2000oC annealed powder sample. The bulk SiBCN1

samples contain much more Si3N4 than the powder analogues, but no crystalline carbon phase

exists after annealing at either temperature.

Considering crystalline Si3N4 formation, the XRD patterns in Figure 41 reveal that bulk

samples always have higher concentrations of Si3N4 than their powder analogues, regardless

of composition. This could be explained if Reaction (2) occurs more easily in the powder

samples. The XRD patterns were measured on polished bulk samples. It demonstrates that

relative to powder samples, the inner part of bulk samples contains more crystalline Si3N4.

Again, this is due to the build-up of N2 pressure within the bulks, impeding the decompostion

of Si3N4. The patterns also reveal that, for both the powder and bulk samples, the boron-

containing materials have higher Si3N4 concentrations than the boron-free samples. This

suggests that boron impedes the degradation of Si3N4.

Several factors can affect the thermal stability of PDCs against crystallization,

including surface area and boron concentration. Surface area effects can be easily seen by

XRD. Regardless of the annealing temperature and composition, powder samples have more

crystalline phases (evidenced by stronger diffraction peaks) than the bulk counterparts,

suggesting that surface area significantly enhances crystallization by providing lower-energy

nucleation sites. On the other hand, comparisons between the SiCN1 and SiBCN1 samples

(b)


reveal that, regardless the sample form, boron-containing samples have more crystalline Si3N4,

indicating that a low concentration of boron impedes Si3N4 degradation.

4.2.3 Summary

Poly(boro)silylcarbodiimide-derived SiCN1 and SiBCN1 ceramics were transformed

into bulk materials via a warm pressing process, and into powder materials by simple thermal

pyrolysis. The results show that the thermal stability of the ceramics strongly depends on the

processing route and on the chemistry of the precursors (i.e., carbon content, precursor type

and presence or absence of boron). The microstructures of SiCN1 and SiBCN1 ceramic

powders pyrolyzed at 1400 ºC were characterized by the presence of α/β-SiC crystallites

embedded in a matrix of amorphous carbon and Si3N4 in contrast to the bulk ceramics which

remain completely amorphous at this temperature. Moreover, the presence of boron in the

SiBCN powder promotes crystallization of SiC. In case of bulk ceramics, SiCN1-1400 shows

a fine organization of carbon. The factors influencing the thermal stability against

crystallization and decomposition are the chemical composition (especially carbon content

and the presence of boron), the architecture of the precursor polymer, and the processing route

in terms of its effect on final residual porosity and surface area.

During annealing at temperatures up to 2100 oC, the crystallization and degradation

behaviour were investigated by elemental analysis, TG/DTA and XRD. Compared with their

bulk analogues, the powder samples exhibited higher weight loss. The bulk samples exhibited

two weight loss events while their powder analogues showed only one, regardless of

composition. The degradation of Si3N4 in the powder samples is mainly related to the

carbothermal reaction, while in the bulk samples both carbothermal reaction and

decomposition occur. This difference is likely due to the greater suface-to-volume ratio of the

powder samples. We also found that the surface area can enhance the crystallization of

ceramics. Samples with larger specific surface areas contain more “easy” sites for the

carbothermal reaction to proceed, and for the crystallization of silicon carbide. Meanwhile,

the presence of boron impedes the degradation of silicon nitride, indicated by the higher Si3N4

concentrations analyzed in boron-containing materials.

Thermodynamic stability 69

4.3 Thermodynamic stability

PDCs exhibit pronounced thermal stability against crystallization and decomposition,

making them excellent candidates for high-temperature and ultra-high-temperature

applications [2,21,105,107,108,156,157]. Many previous works focused on the nanostructure

of these amorphous materials, and furthermore, their thermodynamic stability. It was found

that the energetics of this class of ceramics greatly depends on their structures [10–14]. To

further understand the effects of structure characteristic on the energetics, following two

experiments are designed and carried out in this section:

1) Investigate the effect of the molecular architecture of the precursors, namely linear

[-(PhVi)Si-NCN-]n and branched [-(Ph)Si-(NCN)1.5-]n polymer, on the structure and

energetics of SiCN PDCs. Comprehensive and comparative studies of the structure and

energetics of SiCN ceramics were carried out by using multinuclear (29Si, 13C, 15N, and 1H)

NMR spectroscopy and oxidative drop solution calorimetry. Special emphasis was given to

the previously unexplored aspects of structure and bonding at the interfaces between the

constituent nanodomains and their role in controlling the thermodynamic stability of these

PDCs.

2) Investigate the effect of pyrolysis temperature (1100°C and 1400°C) on the

structural evolution and energetics of amorphous poly(boro)methylvinylsilazane-derived

Si(B)CN PDCs. Similarly, the structural and energetic changes were investigated by solid

state NMR spectroscopy and oxidative solution calorimetry in molten sodium molybdate.

4.3.1 Structure and energetics of polysilylcarbodiimide-derived SiCN ceramics

The structures of ceramics derived from two polysilylcarbodiimides polymers, namely

branched 13C/15N isotope-enriched [-(Ph)Si-(NCN)1.5-]n and linear [-(PhVi)Si-NCN-]n, were

characterized by MAS NMR, and their energetic were measured by high temperature drop

solution calorimetry. The chemical compositions are shown in Table 3 of Chapter 3

(Experimental Section). The hydrogen in these ceramics is residual from the precursor

polymers due to the relatively low pyrolysis temperature used, and the oxygen is from

contamination in air. The nitrogen content is much higher in the Iso_SiCN5 ceramics than in

the SiCN1 ceramics because, in branched 13C/15N isotope-enriched [-(Ph)Si-(NCN)1.5-]n, each

silicon atom connects with a net 1.5 –NCN– group, compared to 1 in [-(PhVi)Si-NCN-]n.

Meanwhile, the carbon content of the SiCN1 ceramics is much higher than the carbon content

of the Iso_SiCN5 ceramics. By increasing the pyrolysis temperature, both polymers lose

hydrogen in the form of hydrogen gas or other organics; however, there is still a noticeable

amount of residual hydrogen in the Iso_SiCN5 ceramic pyrolyzed at 800ºC.

Figure 42 a and b show the 29Si MAS NMR spectra for Iso_SiCN5 and SiCN1

ceramics prepared at 800 and 1100 °C. There is basically only one intense signal at ca. −47

ppm, which is from the SiN4 tetrahedra similar to that in crystalline Si3N4 polymorphs. [158]

70 Thermodynamic stability

Except for SiN4 tetrahedra, other Si tetrahedral environments with the form SiCxN4-x (x=0~4)

are not significant. However, a minor amount of Si−C bonding from SiCN3 tetrahedra cannot

be completely discarded. 29Si nuclides in such environments with 1 C and 3 N nearest

neighbors are expected to have a 29Si chemical shift at about −30 ppm [159,160]. Due to the

absence of the characteristic band at -110 ppm, there appears to be no SiO4 tetrahedra.

Figure 42: 29Si MAS NMR spectra of (a) Iso_SiCN5-800 and Iso_SiCN5-1100 and (b) SiCN1-800 and SiCN1-

1100. Spinning side bands are denoted with asterisks. Reprinted with permission from S. Widgeon, G. Mera, Y.

Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012). Copyright 2012,

American Chemical Society [136].

The 13C CPMAS NMR spectra of the Iso_SiCN5 and SiCN1 ceramics are shown in

Figure 43. The spectra of the Iso_SiCN5 ceramics prepared at both temperatures (800 and

1100 ºC) indicate two isotropic peaks positioned at approximately 126 and 150 ppm (Figure

43 a,b). The chemical shift at 126 ppm is from turbostratic and amorphous carbon, which lie

in the range of 125 to 130 ppm, while the one at ~148 ppm may correspond to sp2C-N

environments [98,161–163]. By carefully fitting the spectra, a calculation was performed to

determine the environment around the carbon atoms. The relative fractions of the two carbon

sites with chemical shift of 126 and 150 ppm are 62 and 38% for Iso_SiCN5-800, and 76 and

24% for Iso_SiCN5-1100, respectively. It is evident that the content of sp2C-C increased with

increasing pyrolysis temperature, while the fraction of sp2C-N decreased. The 13C CPMAS

NMR spectrum of the SiCN1-800 ceramic (Figure 43 c) shows, predominantly, only one

resonance at about 127 ppm. A 13C CPMAS NMR spectrum for the SiCN1-1100 sample was

not acquired due to the insufficiency of hydrogen at 1100ºC.


Figure 43: 13C CPMAS NMR spectra of Iso_SiCN5 ceramics prepared at (a) 800 °C and (b) 1100 °C, and (c)

SiCN1 ceramics prepared at 800 °C. Experimental spectra are plotted as bold solid lines. Simulations of these

spectra are shown with thick dashed lines. Individual simulation components are also shown at the bottom.

Spinning side bands are denoted with asterisks. Reprinted with permission from S. Widgeon, G. Mera, Y. Gao, E.

Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012). Copyright 2012, American

Chemical Society [136].

The 13C MAS NMR spectra of the Iso_SiCN5 and SiCN1 PDC samples are shown in

Figure 44. Besides the dominant resonances at 124 and 150 ppm, there is a third, rather

narrow, resonance centered at 111 ppm that likely corresponds to a small fraction of

graphite [164]. The fitted bands reveal that there are more relative amounts of sp2C-C and

less sp2C-N sites in the Iso_SiCN5-1100 ceramics than in the Iso_SiCN5-800 ceramic. The 13C MAS NMR spectra of the SiCN1 PDC samples pyrolyzed at 800 and 1100 °C (Figure 44

c) display only one broad 13C resonance at about 123 ppm corresponding to an amorphous sp2

C-C environment.

Figure 44: 13C MAS NMR spectra of Iso_SiCN5 PDCs prepared at (a) 800 °C and (b) 1100 °C, and (c) SiCN1

PDCs prepared at 800 °C and at 1100 °C. Experimental spectra are plotted as bold solid lines. Simulations of

the Iso_SiCN5 PDCs spectra are shown with thick dashed lines. Individual simulation components are also

shown at the bottom. Spinning side bands are denoted with asterisks. Reprinted with permission from S.

Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012).

Copyright 2012, American Chemical Society [136].

13C and 15N isotope enrichment of [-(Ph)Si-(NCN)1.5-]n enables one to collect the 15N

MAS and 15N CPMAS NMR spectra from derived Iso_SiCN5-800 and Iso_SiCN5-1100

ceramics (Figure 45). Two Gaussian peaks at 21 and 49 ppm can be simulated from the 15N

MAS and CPMAS NMR spectra of Iso_SiCN5, although only one resonance exists in the 15N

MAS spectrum of Iso_SiCN5-1100. These two bands are assumed to be from the nitrogen

environments in NC1/4Si2/4 and NSi3/4 triangles [162,163,165]. By simulating the direct


polariziation 15N MAS NMR spectra (Figure 45c,d), the relative ratios of these environments

were found to be 18 % and 82 %, respectively, in Iso_SiCN5-800 (Figure 45c).

Figure 45: 15N CPMAS NMR spectra of SiCN5 PDCs prepared (a) at 800 °C and (b) at 1100 °C and 15N MAS

NMR spectra of SiCN5 PDCs prepared at (c) 800 °C and (d) at 1100 °C. Experimental spectra are plotted as

bold solid lines. Simulations of these spectra are shown with thick dashed lines. Individual simulation

components are also shown at the bottom. Reprinted with permission from S. Widgeon, G. Mera, Y. Gao, E.

Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012). Copyright 2012, American

Chemical Society [136].

The 1H MAS NMR spectra of the Iso_SiCN5 ceramic show two resonances centered

at ∼1.5 and 6.5 ppm (Figure 46). The broadness of these two bands indicates that these

hydrogen atoms are bound to the structure and cannot move freely. The two bands at 1.5 and

6.5 ppm are due to the N-H and sp2C-H chemical bonds, respectively. The line shapes can be

simulated using 50% Gaussian - 50% Lorentzian peaks. For ceramics pyrolyzed at either

temperature, the relative ratio of N-H and sp2C-H is about 65:35.


Figure 46: 1H MAS NMR spectra of Iso_SiCN5 PDCs at 800 °C and 1100 °C shown on the same intensity scale. The spectrum of the Iso_SiCN5-800 ceramic has higher intensity than that of the Iso_SiCN5-1100 ceramic indicating loss of hydrogen that occurs during pyrolysis between 800 and 1100 °C. Reprinted with permission from S. Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012). Copyright 2012, American Chemical Society [136].

Considering all information, these polysilylcarbodiimide-derived SiCN ceramics

primarily have only one kind of silicon environment, i.e., the SiN4 tetrahedra which

characterize the Si3N4 nanodomains, as indicated by the 29Si MAS NMR spectra. In the

SiCN1 ceramics, carbon exists basically as amorphous sp2C-C bonds, based on the 13C MAS

and CPMAS NMR spectra. This means that, for the SiCN1 ceramic, the volume fraction of

the “free” carbon domains is significantly larger than the interfacial region. The “free” carbon

probably forms a continuously interconnected amorphous carbon matrix in the structure, with

Si3N4 nanodomains embedded in this matrix. The Iso_SiCN5-800 ceramic has sp2C-N bonds

in addition to the primary sp2C-C bonds. The sp2C-N bonds are likely positioned at the

interface between the Si3N4 nanodomains and the “free” carbon regions in the structure. The 15N MAS NMR spectra of the Iso_SiCN5 ceramics also show that the majority of the N atoms

are in NSi3 sites in Si3N4 nanodomains, while a relatively minor fraction of the N atoms are

bonded to 2 Si and 1 C nearest neighbors, and therefore correspond to the atomic

configurations at the interfacial regions between the Si3N4 and the C nanodomains. This

finding aligns with the 13C MAS NMR results. The C−N−Si linkages are also responsible for

the 13C NMR resonance at about 150 ppm, which represents C-N bonds. The intensities of

both the 13C NMR resonance at ∼150 ppm and the 15N NMR resonance at ∼24 ppm are

preferentially enhanced upon cross-polarization with 1H (Figure 43 a,b and Figure 45 a,b).

These results imply a heterogeneous distribution of protons in the Iso_SiCN5 ceramic. The

protons are in much larger abundances near the NC1Si2 sites in the interfacial regions than in

the interiors of the Si3N4 and the “free” C nanodomains. For the Iso_SiCN5-1100 ceramic,

there is no observable signal from the NC1Si2 environment in the 15N MAS NMR spectrum,

which is consistent with the strong drop in intensity of the resonance at 150 ppm in the 13C

MAS NMR spectrum. It is believed that the concentration of the NC1Si environments at the

interfacial region between the Si3N4 and the “free” C nanodomains decreases with increasing

pyrolysis temperature. Concomitantly the hydrogen concentration also decreases. According

to the relative fractions of N and H atoms at the interfacial NC1Si2 environments, and the total

N and H contents in the Iso_SiCN5-800 ceramic (Table 3 in chapter 3), there are nearly 3 H


atoms per NC1Si2 site. The large loss of hydrogen with an increase pyrolysis temperature

leads to a more complete separation of the Si3N4 and C nanodomains in Iso_SiCN5-1100

ceramic. This process can be accomplished via bond switching, as shown schematically in

Figure 47. Figure 48a shows a schematic of the Iso_SiCN5 ceramic structure with coexisting

a-Si3N4 and amorphous carbon domains, and an interfacial region constructed by NC1Si2 sites.

In the case of the SiCN1 PDCs, as discussed above, the carbon domain is possibly continuous,

forming a matrix in which the a-Si3N4 nanodomains are embedded (Figure 48b).

Figure 47: Schematic representation of (left) mixed bonding between Si, C, N and H atoms in the interfacial

region between Si3N4 and C nanodomains in Iso_SiCN5-800 and (right) loss of hydrogen and of mixed bonding

upon pyrolysis at 1100 ⁰C in Iso_SiCN5-1100 facilitated by local bonding switching. Reprinted with permission

from S. Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181

(2012). Copyright 2012, American Chemical Society [136].

Figure 48: Cartoons of the nanodomain structural models for (a) Iso_SiCN5 and (b) SiCN1 PDCs. The

nanometer-scale amorphous/turbostratic carbon and Si3N4 domains are shown with stripes and dots, respectively,

with the interfacial areas as the white regions. (b) shows a continuous matrix of amorphous/turbostratic carbon

with Si3N4 clusters embedded in it. The atomic structures within these domains are shown schematically in the

insets. Reprinted with permission from S. Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R.

Riedel, Chem. Mater. 24, 1181 (2012). Copyright 2012, American Chemical Society [136].

The high temperature drop solution calorimetry experiments measure the heat of the

reaction shown in Scheme 16,

SivCwNxOyHz (solid, 25 oC) + (v + w - y/2 + z/4) O2 (gas, 809 oC) = vSiO2 (cristobalite, 809 oC) + wCO2

(gas, 809 oC) + x/2 N2 (gas, 809 oC) + z/2 H2O (gas, 809 oC)

Scheme 16

where the values of the coefficients v, w, x, y, and z are given in Table 7.


Table 7 Enthalpies of drop solution (Hds), enthalpies of oxidation at 25 oC (Hox), enthalpies of formation from the elements (H0

f, elem), and enthalpies of formation from the compounds (see text for details) (H0f, comp). Uncertainty

is two standard deviations of the mean. Numbers in parentheses are numbers of calorimetry drops for each sample.

Sample g-atom (v + w + x + y + z = 1) Enthalpies at 25 oC (kJ/g-atom)

Siv Cw Nx Oy Hz Hds Hox Hf, elem Hf, comp

Iso_SiCN5-800 0.096 0.346 0.244 0.041 0.273 -161.1 ± 1.2 (9) -179.6 ± 1.2 -82.6 ± 1.2 -42.5 ± 1.2

Iso_SiCN5-1100 0.134 0.435 0.285 0.015 0.130 -226.6 ± 1.4 (10) -243.0 ± 1.4 -68.4 ± 1.4 -25.8 ± 1.5

SiCN1-800 0.121 0.594 0.149 0.020 0.116 -272.0 ± 1.2 (8) -287.9 ± 1.2 -72.7 ± 1.3 -32.0 ± 1.3

SiCN1-1100 0.126 0.671 0.151 0.019 0.034 -302.0 ± 1.7 (9) -316.5 ± 1.7 -66.5 ± 1.8 -25.9 ± 1.8

The heat of drop solution, ΔHds, is the sum of the heat pick-up of the sample from

room temperature to the calorimeter temperature, and the heat of the anticipated reaction.

From the experimental data (Table 7), the enthalpies of drop solution are strongly negative

due to the dominant energetic effect of C and N oxidation. The CO2 and N2 gases released

after oxidation are swept from the headspace above the solvent using continuously flowing

oxygen. Using the thermodynamic data from Table 8, Step 1, the enthalpy of oxidation, ΔHox,

at 25 °C for the two kinds of ceramics can be calculated using the reaction shown in Scheme

17.

ΔH0ox = ΔHds - vΔH2 + (v+w-y/2+z/4) ΔH3 - wΔH4 - x/2 ΔH5 - z/2 ΔH6

Scheme 17

The reference heat contents of SiO2, O2, CO2, N2, and H2O are marked ΔH2, ΔH3, ΔH4,

ΔH5, and ΔH6, respectively [166]. The enthalpy of formation from the elements ΔHf, elem at

25 °C for the two ceramics follow from their enthalpy of oxidation ΔHox (Step 2 in Table 8),

and can be generalized by Scheme 18,

ΔH0f, elem = -ΔHox + vΔHf(SiO2) + wΔHf(CO2) + z/2 ΔHf(H2O)

Scheme 18

where the values of ΔHf are listed in Table 7, and refer to the enthalpies of formation of

cristobalite, carbon dioxide, and water from the elements. [167]

In order to calculate the difference in enthalpy between of the amorphous PDCs and an

isocompositional mixture of the binary compounds, it is assumed that all of the oxygen in the

samples is bonded to silicon forming SiO2, and the remaining silicon is bonded to nitrogen in

form of Si3N4. The “free” carbon is assumed to be in form of graphite.

For the Iso_SiCN5 ceramics, “extra” nitrogen is assumed to be N2 gas in the

compositional mixture. In this case, the difference in the enthalpy of formation the PDCs from

the binary compounds is calculated using the reaction in Scheme 19.

ΔHf, SiCNO = ΔH11 - y/2 ΔH12 - (v-y/2)/3 ΔH13

Scheme 19


In the case of the SiCN1 samples, there is sufficient Si to bond to all of the nitrogen,

so that nitrogen is not counted as separate component. Thus, the enthalpy differences are

calculated using reaction shown in Scheme 20.

ΔHf, SiCNO = ΔH11 - y/2 ΔH12 - x/4 ΔH13

Scheme 20

The enthalpies, ΔHf, comp, are calculated using the thermodynamic cycles. The

measured enthalpies of drop solution, ΔHds, and the calculated values for ΔHox, ΔHf, elem, and

ΔHf, comp are reported in Table 7.

Table 8. Thermodynamic cycles for SiCN1 and Iso_SiCN5 ceramics

Step 1: Enthalpies of oxidation (ΔH0ox) at 25 oC:

SivCwNxOyHz (solid, 25 oC) + (v+w-y/2+z/4) O2 (gas, 809 oC) → v SiO2 (cristobalite, 809 oC) + w CO2 (gas, 809 oC) + x/2 N2 (gas, 809 oC) + z/2 H2O (gas, 809 oC)

ΔH1 = ΔHds of SiCN-PDCs

SiO2 (cristobalite, 25 oC) → SiO2 (cristobalite, 809 oC) ΔH2 = 50.48 kJ/mol [166]

O2 (gas, 25 oC) → O2 (gas, 809 oC) ΔH3 = 25.51 kJ/mol [166]

CO2 (gas, 25 oC) → CO2 (gas, 809 oC) ΔH4 = 37.78 kJ/mol [166]

N2 (gas, 25 oC) → N2 (gas, 809 oC) ΔH5 = 24.77 kJ/mol [166]

H2O (liquid, 25 oC) → H2O (gas, 809 oC) ΔH6 = 73.6 kJ/mol [166] SivCwNxOyHz (solid, 25 oC) + (v+w-y/2+z/4) O2 (gas, 25 oC) → v SiO2 (cristobalite, 25 oC) + w CO2 (gas, 25 oC) + x/2 N2 (gas, 25 oC) + z/2 H2O (liquid, 25 oC)

ΔH0ox

ΔH0ox = ΔH1 - vΔH2 + (v+w-y/2+z/4)ΔH3 - wΔH4 - x/2ΔH5 - z/2 ΔH6

Step 2: Enthalpies of formation from the elements (ΔH0f, elem) at 25 oC:

SivCwNxOyHz (solid, 25 oC) + (v+w-y/2+z/4) O2 (gas, 25 oC) → v SiO2 (cristobalite, 25 oC) + w CO2 (gas, 25 oC) + x/2 N2 (gas, 25 oC) + z/2 H2O (liquid, 25 oC)

ΔH7 = ΔH0ox

Si (solid, 25 oC) + O2 (gas, 25 oC) → SiO2 (cristobalite, 25 oC) ΔH8 = -908.346 ± 2.1 kJ/mol [167]

C (solid, 25 oC) + O2 (gas, 25 oC) → CO2 (gas, 25 oC) ΔH9 = -393.522 ± 0.05 kJ/mol [167]

H2 (gas, 25 oC) + 1/2 O2 (gas, 25 oC) → H2O (liquid, 25 oC) ΔH10 = -285.83 ± 0.042 kJ/mol [167]

v Si (solid, 25 oC) + y/2 O2 (gas, 25 oC) + x/2 N2 (gas, 25 oC) + w C (solid, 25 oC) + z/2 H2 (gas, 25 oC) → SivCwNxOyHz (solid, 25 oC)

ΔH0f, elem

ΔH0f, elem = -ΔH7 + vΔH8 + wΔH9 + z/2 ΔH10

Step 3: Enthalpies of formation from the compounds at 25 oC: v Si (solid, 25 oC) +y/2 O2 (gas, 25 oC) + x/2 N2 (gas, 25 oC) + w C (solid, 25 oC) + z/2 H2 (gas, 25 oC) → SivCwNxOyHz (solid, 25 oC)

ΔH11 = ΔH0f, elem

Si (solid, 25 oC) + O2 (gas, 25 oC) → SiO2 (cristobalite, 25 oC) ΔH12 = -908.346 ± 2.1 kJ/mol [167]

3Si (solid, 25 oC) +2N2 (gas, 25 oC) → Si3N4 (solid, 25 oC) ΔH13 = -850.20 ± 2.7 kJ/mol [95]

Iso_SiCN5 samples: from SiO2 (cristobalite), Si3N4, N2, H2 and C (graphite) (ΔH0f, comp)at 25 oC

y/2 SiO2 (cristobalite, 25 oC) + (v-y/2)/3 Si3N4 (solid, 25 oC) + + [x-[4/3(v-y/2)]]/2 N2 (gas, 25 oC) + w C (graphite, 25 oC) + z/2 H2 (gas, 25 oC) → SivCwNxOyHz (solid, 25 oC) ΔHf, comp = ΔH11 - y/2 ΔH12 - (v-y/2)/3 ΔH13

ΔHf, comp


SiCN1 samples: from SiO2 (cristobalite), Si3N4, H2 and C (graphite) (ΔH0f, comp) at 25 oC

y/2 SiO2 (cristobalite, 25 oC) + x/4 Si3N4 (solid, 25 oC) + w C (graphite, 25 oC) + z/2 H2 (gas, 25 oC)→ SivCwNxOyHz (solid, 25 oC) ΔHf, comp = ΔH11 - y/2 ΔH12 - x/4 ΔH13

ΔHf, comp

As shown in Table 7, the thermodynamic analysis shows that all samples have

negative enthalpies of formation from the components, and the ceramics pyrolyzed at 800 °C

are more energetically stable than the ceramics pyrolyzed at 1100 °C. Also, the Iso_SiCN5-

800 sample is somewhat more energetically stable than the SiCN1-800 PDC. Considering a

recently-published study on other phenyl-containing polysilylcarbodiimides-derived

ceramics [10], the materials in this work are energetically more stable than those of the

previous study. The main difference between the previously studied SiCN samples and the

materials investigated in this work is that the precursors in the former study contain -H, -

phenyl or –methyl functional groups that form a linear chain. Here, while the SiCN1

precursor contains vinyl groups and is also a linear polymer, the Iso_SiCN5 preceramic

polymer is branched. Moreover, the ceramics in the previous study have small amounts of

SiC [10], whereas the ceramics in this work contain no SiC component detectable by NMR.

As discussed previously, the 13C MAS NMR data indicate that, compared to the

SiCN1 ceramics, the Iso_SiCN5 ceramic prepared at 800 °C contains more mixed N and C

bonding (i.e., in the NC1Si2 environments at the interface between Si3N4 and sp2 carbon

nanodomains). The higher concentration of mixed bonds may also be due to the fact that the

Iso_SiCN5 ceramic was derived from a branched preceramic polymer which contains a larger

percentage of nitrogen. These N atoms are connected to sp2-carbon. Combined with the

calorimetry results, the higher concentration of mixed bonds between C, N, and Si atoms in

the Iso_SiCN5-800 ceramic contributes to its greater energetic stability than Iso_SiCN5-1100.

It can be concluded that mixed bonding in interfacial regions stabilize the structure of these

PDCs. This hypothesis is consistent with the calorimetry results of Iso_SiCN5-1100 and

SiCN1-1100 ceramics, which have no considerable hydrogen at the interfacial region. These

materials have identical energetics within experimental error, and are less negative (less stable)

than their counterparts pyrolyzed at 800 °C. Moreover, similar observations have also been

made in previous studies on SiCO PDCs based on ab initio calculations that indicated that the

presence of mixed Si−C−O bonds energetically stabilized the structures of these

materials [96].

It is interesting to argue that, the energetic stabilization of mixed bonding coupled with

its diminution at higher temperature is a thermodynamic contradiction. There must be a

decrease in entropy along with the disappearance of mixed bonding and the coarsening and

ordering of Si3N4 and C domains. This process is associated with a positive enthalpy change.

As the calorimetry results suggest, the free energy of such a process would be positive too.

Therefore there would be no driving force for the demixing of mixed bonds. However,

hydrogen evolution also results in a change in entropy. The evolution of hydrogen gas and the


disruption of the H-stabilized mixed bonding environments is then associated with positive

ΔH and positive ΔS terms, and is favored with increasing temperature. The energetic

stabilization of the mixed bonds may require the presence of hydrogen near the NC1Si2

environments. The positioning of hydrogen at mixed bonds in the interfacial region can be

interpreted as stabilizing to the structure and energetic of these PDCs.

4.3.2 Structure and energetics of poly(boro)silazane-derived Si(B)CN ceramics

In this section, the effect of pyrolysis temperature (1100 °C and 1400 °C) on the

structural evolution and energetics of amorphous poly(boro)silazane-derived SiCN4 and

SiBCN4 PDCs is systematically investigated by solid state NMR spectroscopy and oxidative

solution calorimetry in molten sodium molybdate.

Chemical compositions of the samples pyrolyzed at 1100 and 1400 °C, both in weight

and atom percentages, are given in Table 9. High temperature thermogravimetric

measurements of SiCN4-1100 and SiBCN4-1100 (Figure 49a) show less than 0.5 % mass loss

from 1100 to 1400 °C, confirming that for both the boron-containing and boron-free samples,

the amounts of the main elements, namely Si, N, C, and B (in case of boron containing-

samples), remain constant with increasing pyrolysis temperature in this range. Similarly to

most Si-(B-)C-N PDCs, SiCN4 and SiBCN4 ceramics pyrolyzed at 1100 °C are X-ray

amorphous (Figure 49b) [21,105,136]. After heat treatment at 1400 ºC, the XRD patterns of

these samples are characterized by a very broad reflection extended over the 2θ range of 30 to

50 degrees that may be caused by a change in medium-range order, for example involving the

clustering of SiC4 tetrahedra.

Table 9 Chemical compositions of SiCN4-1100, SiCN4-1400, SiBCN4-1100, and SiBCN4B-1400.

Sample N

wt%* O

wt% C

wt% Si

wt% B

wt% H

wt% Cl

wt%

Molecular Weight

(g/g-atom)

SiCN4- 1100

17.10 (20.49)

1.56 (1.64)

29.19 (40.81)

51.75 (31.01)

— 0.36

(6.04) 0.04

(0.02) 16.78

SiCN4- 1400

17.23 (21.20)

1.44 (1.55)

29.84 (42.84)

51.283 (31.55)

— 0.165 (2.84)

0.042 (0.02)

17.23

SiBCN4- 1100

13.60 (16.71)

4.63 (4.98)

28.85 (41.35)

48.36 (29.70)

4.52 (7.07)

0.01 (0.17)

0.03 (0.01)

17.20

SiBCN4- 1400

14.09 (17.28)

4.59 (4.93)

28.39 (40.62)

48.20 (29.55)

4.68 (7.44)

0.01 (0.17)

0.04 (0.02)

17.17

*at. % in parentheses.


1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

75

80

85

90

95

100

SiBCN4 - 1100

Mas

s (%

)

Temperature (C)

SiCN4 - 1100

10 20 30 40 50 60 70 80 90

SiCN4 - 1100

SiBCN4 - 1400

SiCN4 - 1400

SiBCN4 - 1100

Inte

nsi

ty (

a.u

.)

()

Figure 49 Characterization of the ceramic samples: (a) high temperature thermogravimetric analysis of SiCN4-1100 and SiBCN4-1100, (b) XRD patterns of SiCN4-1100, SiCN4-1400, SiBCN4-1100, and SiBCN4-1400. Reprinted from Scripta Materialia, V 69, Y. Gao, S. J. Widgeon, T. B. Tran, A. H. Tavakoli, G. Mera, S. Sen, R. Riedel, and A. Navrotsky, Effect of Demixing and Coarsening on the Energetics of Poly(boro)silazane-Derived Amorphous Si-(B-)C-N Ceramics, 347, copyright (2013), with permission from Elsevier [135].

Solid-state MAS NMR provides high-resolution structural information on amorphous

materials that cannot be gleaned from XRD. The 29Si MAS NMR spectra of boron free

samples (SiCN4, Figure 50a) show a broad resonance that spans from ca. -70 to +10 ppm that

encompasses the characteristic chemical shifts of SiN4, SiCN3, SiC2N2, SiC3N, and SiC4,

species located at -48, -35, -5, +3 and -18 ppm, respectively. The relative intensity of the

SiCN3 peak, located at ~ -35 ppm, decreased, and concomitantly, the intensity of the SiC4

peak increased with increasing pyrolysis temperature, suggesting that mixed SiCxN4-x bonds

undergo demixing to form additional SiC4 and SiN4 species. The 13C MAS NMR spectrum

(Figure 50b) shows resonances around 140 and 122 ppm that correspond to sp2 C-N and sp2

C-C environments, respectively [131]. As the pyrolysis temperature is increased from 1100 to

1400 oC, the relative amount of C-N bonds decreases, and the relative amount of C-C bonds

increases, again indicating the demixing of C and N and the release of low amounts of

nitrogen as a gaseous by-product. As observed in previous studies, C-N bonds only exist as

the interface between SiCN and carbon domains [101]. A decrease in the relative amount of

C-N bonds suggests a decrease in the relative mass fraction of interdomain boundaries, and

when taken together with the increase in the relative amount of C-C bonds between heat

treatments at 1100 and 1400 ºC, strongly suggests the coarsening of carbon domains.

40 20 0 -20 -40 -60 -80 -100 -120


SiC4SiCN3

SiN4

300 250 200 150 100 50 0 -50 -100


sp2 Csp2 C - N

sp3 C - Si

(a) (b)

(a) (b)


Figure 50 (a) 29Si and (b) 13C MAS NMR of SiCN4-1100 (thin line) and SiCN4-1400 (bold line). Reprinted from Scripta Materialia, V 69, Y. Gao, S. J. Widgeon, T. B. Tran, A. H. Tavakoli, G. Mera, S. Sen, R. Riedel, and A. Navrotsky, Effect of Demixing and Coarsening on the Energetics of Poly(boro)silazane-Derived Amorphous Si-(B-)C-N Ceramics, 347, copyright (2013), with permission from Elsevier [135].

The 29Si MAS NMR spectra (Figure 51a) of the boron-containing SiBCN4 samples

also show a demixing of N and C, resulting in a reduction in the relative concentration of the

SiC4N4-x(x=0-4) mixed bonding environments with increasing pyrolysis temperature. This

process is manifested in Figure 51a by an increase in the relative amount of SiC4 tetrahedral

environments with respect to the SiCN3 environments. Since the TG analysis does not show

mass loss after heating to 1400 ºC, the concentrations of Si and N remain constant, and the

demixing of N and C therefore implies that, with increasing temperature, Si-N bonds break,

and new B-N bonds are formed. This hypothesis is supported by the 11B MAS NMR spectra

(Figure 51b) showing that, with increasing pyrolysis temperature, the relative amount of

BCN2 environments (~ 40 ppm) decreases, while that of the BN3 environments (~30 ppm)

increases. Here, the C lost from the BCN2 environments contributes to the increase in sp2 C-C

bonds in carbon domains. This is directly observed in the 13C MAS NMR spectra (Figure 51c)

where, similar to boron-free SiCN4 sample, the relative amount of C=C bonds increases with

increasing pyrolysis temperature. Hence, these data show a decrease in the SiCN-carbon

interdomain region. In addition, the 13C MAS NMR spectra (Figure 51c) indicate a

tremendous increase in Si-C(sp3) bonds (~28 ppm), providing further proof that pyrolysis at

1400 °C results in more SiC4 tetrahedral units than at 1100 °C.

40 20 0 -20 -40 -60 -80 -100 -120


SiC4SiCN3

SiN4

70 60 50 40 30 20 10 0 -10 -20


BN3

BCN2

300 250 200 150 100 50 0 -50 -100

sp2 C - B


sp2 Csp2 C - N sp3 C - Si

Figure 51 (a) 29Si, (b) 11B, and (c) 13C MAS-NMR of SiBCN4-1100 (thin line) and SiBCN4-1400 (bold line). * Reprinted from Scripta Materialia, V 69, Y. Gao, S. J. Widgeon, T. B. Tran, A. H. Tavakoli, G. Mera, S. Sen, R. Riedel, and A. Navrotsky, Effect of Demixing and Coarsening on the Energetics of Poly(boro)silazane-Derived Amorphous Si-(B-)C-N Ceramics, 347, copyright (2013), with permission from Elsevier [135].

The drop solution enthalpies, ΔHds, of SiCN4 and SiBCN4 corresponding to reactions

(1,2,4,5) in Table 10 are listed in the same table. As previously discussed, the compositions of

SiCN4 and SiBCN4 are the same after pyrolysis at 1100 and 1400 °C. Thus, directly

comparing drop solution enthalpies using the thermodynamic cycles given in Table 10 yields

the enthalpy of structural evolution, ΔHS-E. The exothermic ΔHS-E values of –13.5 ± 4.7 and –

8.8 ± 5.7 kJ/g-atom obtained, respectively, for SiCN4 and SiBCN4 reveal that the ceramics

heat treated at 1400 °C are energetically more stable than those pyrolyzed at 1100 °C.

(a) (b) (c)


Table 10 Thermodynamic cycles for calculating the enthalpies of structural evolution, ΔHS-E, of samples SiCN4 and SiBCN4 between 1100 and 1400 °C.

Reaction Enthalpy

(kJ/g-atom)

a. Enthalpy of the structural evolution of SiCN4 between 1100 and 1400 °C (HS-

E(SiCN4))

(1) SiaCbNcOdHe (SiCN4-1100), 25°C + (a+b-d/2+e/4+3c/4) O2 (gas, 802⁰C) → a SiO2 (cristobalite)* + b CO2 (gas) + c/2 N2 (gas) + e/2 H2O (gas)

H1 = –348.6 ± 3.5

(2) SiaCbNcOdHe (SiCN4-1400) + (a+b-d/2+e/4+3c/4) O2 (gas, 802⁰C) → a SiO2 (cristobalite) + b CO2 (gas) + c/2 N2 (gas) + e/2 H2O (gas)

H2 = –335.1 ± 3.2

(3) SiaCbNcOdHe (SiCN4-1100) → SiaCbNcOdHe (SiCN4-1400) H3 = HS-E(SiCN4)

H3 = HS-E(SiCN4) = H1 – H2 = –13.5 ± 4.7 kJ/g-atom

b. Enthalpy of the structural evolution of SiBCN4 between 1100 and 1400 °C (HS-

E(SiBCN4))

(4) SivCwNxOyBz (SiBCN4-1100), 25⁰C + (v+w-y/2+3z/4) O2 (gas, 802⁰C) → v SiO2 (cristobalite) + w CO2 (gas) + x/2 N2 (gas) + z/2 B2O3

H4 = –322.3 ± 4.2

(5) SivCwNxOyBz (SiBCN4-1400), 25⁰C + (v+w-y/2+3z/4) O2 (gas, 802⁰C) → v SiO2 (cristobalite) + w CO2 (gas) + x/2 N2 (gas) + z/2 B2O3

H5 = –313.5 ± 3.9

(6) SivCwNxOyBz (SiBCN4-1100) → SivCwNxOyBz (SiCN4-1400) H6 = HS-E(SiBCN4)

H6 = HS-E(SiBCN4) = H4 – H5 = –8.8 ± 5.7 kJ/g-atom * The temperature of products in reactions (1,2,4,5) is 802°C.

In SiBCN4 ceramic prepared at 1100 C, the presence of only BN2C sites, combined

with the lack of BNC2 and BC3 sites, indicates that the BN units are likely interfacial. It is

suggested that the majority of the BN triangles must be bonded to SiCxN4-x (x=0~4) domains

and carbon domains through Si-C-B and B-C-C (or N-C-C) bridges, respectively. When the

pyrolysis temperature is increased to 1400 C, Si-N bonds break, and B-N and Si-C bonds

form. Meanwhile, the cleavage of interfacial B-C and N-C bonds leads to the formation of B-

N and C-C bonds. Similarly, in the case of SiCN4 ceramic, the cleavage of C-N bonds at the

interfacial regions between SiCxN4-x (x=0~4) and carbon domains occurs with increasing

pyrolysis temperature. The breaking of bonds at the nanodomain interfaces results in an

increase of the interface energy, providing a driving force for further nanodomain coarsening

in order to compensate the increased energy of the system caused by the missing interdomain

bonds. Therefore, demixing of the interfacial bonds can energetically promote coarsening of

the SiCxN4-x and carbon nanodomains, as suggested by the NMR analysis.

We conclude that demixing and coarsening, occurring concurrently, stabilize the PDC

structures. It is obvious that the coarsening processes reduce the total energy of a system

through the elimination of interfaces. A previous NMR study on carbon-rich Si-C-N PDCs

indicates that the presence of mixed bonded SiCxN4-x (x=0-4) in the interfacial region between

amorphous sp2 C and silicon nitride may contribute to thermodynamic stability only if local H

atoms are present at the interface, as discussed in Section 4.3.1 (also see [136]). In the present

study, Si-(B-)C-N ceramics were derived from a different precursor, poly(boro)silazane, and

were pyrolyzed at higher temperatures. Since the hydrogen content in these samples is


substantially lower, it is less likely to play a significant role in the structural or

thermodynamic stability of the ceramics.

4.3.3 Summary

Multinuclear MAS NMR spectroscopy demonstrated that the structure of carbon rich

Iso_SiCN5 and SiCN1 SiCN ceramics derived from phenyl-containing polysilylcarbodiimides

consists predominantly of separate amorphous silicon nitride and sp2 carbon nanodomains.

The results of oxide melt solution calorimetry indicate that both Iso_SiCN5 and SiCN1

ceramics are energetically more stable after pyrolysis at 800 °C than after pyrolysis at 1100°C.

Moreover, Iso_SiCN5 ceramics are also more energetically stable than SiCN1 ceramics. The

energetic stabilization of the Iso_SiCN5 ceramic is attributed to the presence of mixed

bonding between N, C, and Si atoms at the interfacial region between the Si3N4 and C

nanodomains, which could be attributed to the nature of their precursors: Iso_SiCN5 ceramics

are derived from a branched polysilsesquicarbodiimide, while SiCN1 is formed from a linear

polysilylcarbodiimide. The relative concentrations of the mixed bonds between N, C, and Si

atoms decrease upon an increase in the pyrolysis temperature to 1100 °C, along with a

concomitant loss of hydrogen. It is hypothesized that the mixed bonds are stabilized by

hydrogen bonding at the interfacial regions between the Si3N4 and C nanodomains, and that

the loss of this hydrogen at higher temperatures provides the driving force for the destruction

of mixed bonding.

The effect of pyrolysis temperature on the structural evolution of amorphous

poly(boro)methylvinylsilazane-derived Si-(B-)C-N was investigated. By MAS NMR, the

main processes governing structural evolution between 1100 and 1400 ºC are identified as (i)

the demixing of SiCxN4-x(x=0-4) mixed bonding environments, (ii) the cleavage of mixed

bonds at interdomain regions, and (iii) the coarsening of domains. Calorimetry demonstrates

that this structural evolution is favorable in both enthalpy and free energy.

Solid state structure and microstructure of Si(B)CN ceramics 83

4.4 Solid state structure and microstructure of Si(B)CN ceramics

The existence of nanodomains in X-ray amorphous PDCs has attracted extensive

attention, because nanodomains are considered to be responsible for the thermal stability,

crystalline behavior, and other properties of the ceramics. In this section, we investigate the

chemical environment of nanodomains and nanodomain interfaces, and the nanodomain size

of the ceramics. The ceramics are characterized by (i) MAS NMR, which can probe chemical

bonding information to reveal the local environment of atoms; (ii) Small angle X-ray

scattering (SAXS) which can be used to determine the average nanodomain size. MAS NMR

and SAXS are both integral methods and can reveal the overall structural information of the

materials. This investigation is assisted by other techniques, including XRD, TEM and Raman

spectroscopy. The effect of boron modification, carbon content, and precursor molecular

architecture on the final nanodomain structure is discussed.

4.4.1 Solid state structure of Si(B)CN ceramics

Structural investigations by MAS NMR spectroscopy show that the segregation of N

and C into separate Si tetrahedra is a characteristic of polysilylcarbodiimide-derived SiCN

ceramics. This is in strong contrast to the observation of significant N and C mixing in the

nearest neighbor coordination environment of Si in polysilazane-derived SiCN PDCs. The

structure of nanodomain and nanodomain interfaces of the ceramics derived from tailored

precursors are characterized by 29Si,13C and 11B MAS NMR. 29Si and 13C MAS NMR spectra

(Figure 52 and Figure 53) were measured on SiCN ceramics. 11B MAS NMR (Figure 54) was

additionally completed on the boron-modified SiBCN ceramics. The selected ceramics were

prepared at 1100°C and 1400°C. 1100°C is the standard temperature for producing amorphous

PDCs, and 1400°C is a temperature at which phase separation occurs, but no deleterious

crystallization initiates.


40 20 0 -20 -40 -60 -80 -100 -12029Si chemical shift (ppm)

SiCN1-1100 SiCN1-1400


SiBCN1-1100 SiBCN1-1400


SiCN2-1100 SiCN2-1400



Figure 52 29Si MAS NMR spectra (to be continued)

(a)

(b)

(c)

(d)


40 20 0 -20 -40 -60 -80 -100 -120


SiCN3-1100 SiCN3-1400




SiCN4-1100 SiCN4-1400



(f)(h)

(e)(g)

(continued) 29Si MAS NMR


300 250 200 150 100 50 0 -50


SiCN1-1100

300 250 200 150 100 50 0 -5013C chemical shift (ppm)


300 250 200 150 100 50 0 -50


SiCN2-1100 SiCN2-1400



Figure 53 13C MAS NMR spectra (to be continued)

(a)

(b)

(c)

(d)


300 250 200 150 100 50 0 -50


SiCN3-1100 SiCN3-1400

300 250 200 150 100 50 0 -50


SiBCN3-1100


SiCN4-1100 SiCN4-1400



(e)

(f)

(g)

(h)

(continued) 13C MAS NMR


120 100 80 60 40 20 0 -20 -40 -60 -80 -100



120 100 80 60 40 20 0 -20 -40 -60 -80 -100



120 100 80 60 40 20 0 -20 -40 -60 -80 -100



120 100 80 60 40 20 0 -20 -40 -60 -80 -100



Figure 54 11B MAS NMR spectra

(b)

(c)

(d)

(a)


4.4.1.1 Solid state structure of polysilylcarbodiimide-derived SiCN1and SiCN2 ceramics

The chemical bonding environments of silicon atoms and carbon atoms were obtained

by MAS NMR for polylsilylcarbodiimide-derived SiCN ceramics, i.e. SiCN1, SiCN2

ceramics. The separation of the Si3N4 and “free” carbon nanodomains was seen in previous

structural studies of SiCN1 ceramic [136]. The 29Si MAS NMR spectra (Figure 52 (a)) of

these ceramics pyrolyzed at 1100ºC show mainly one heterogeneously broadened Gaussian

peak with a 29Si isotropic chemical shift (δiso) of ~ -48ppm, corresponding to the presence of

primarily one kind of silicon environment i.e. the SiN4 tetrahedral units that are characteristic

of Si3N4 nanodomains. This result implies almost complete separation of nitrogen and carbon,

and the formation of Si3N4 and “free” carbon nanodomains in the structures of these PDCs.

However, some Si−C bonding may be present at the interfaces of these domains. The 13C

MAS NMR spectrum of the SiCN1-1100 PDC (Figure 53(a)) contains primarily one isotropic

peak at δiso ~125 ppm, characteristic of sp2 hybridized carbon network, and the broad nature

of the peak indicates an amorphous structure. This means C is present primarily as amorphous

sp2-bonded carbon domains. In the 13C NMR spectra of SiCN1 PDCs, the absence of any

significant NMR signal near ~150 ppm, corresponding to C−N bonds, indicates a significantly

smaller fraction of such carbon environments in the structure compared to that of “free”

carbon. Consequently, this result also implies that the average size of the “free” carbon

domains is very large in highly carbon-rich SiCN1 PDC. The “free” carbon possibly forms a

continuously interconnected amorphous carbon matrix in this kind of PDC. In this case, the

Si3N4 nanodomains are expected to be embedded in the amorphous carbon matrix in the case

of the SiCN1 PDCs (Figure 55(a)).

SiCN2 ceramics were derived from [-(MeVi)Si-NCN-]n, which has a methyl group

and lower carbon content compared to the SiCN1 precursor. The 29Si MAS NMR spectrum of

SiCN2-1100 (Figure 52(b)) primarily contains SiN4 tetrahedral units and SiN3(NCN) units.

The signal of the SiN3(NCN) units are centered at ca. -60 ppm. FT-IR spectrum of SiCN2-

1100 confirms the presence of –N=C=N- group which has the signal at ca. 2150cm-1 (Figure

56). According to the 13C MAS NMR spectrum (Figure 53(b)), there is a significant amount

of sp2 C-N bonds in the structure of the SiCN2-1100 ceramic, which is different from that of

SiCN1. These sp2 C-N bonds are located at the interface between the SiN4 and the “free”

carbon nanodomains, and no considerable amount of Si-C bonding is observed. A schematic

is shown in Figure 55(b).

90 Solid state structure and microstructure of Si(B)CN ceramics

(a)

(b)

Figure 55 Schematic drawing of polysilylcarbodiimide-derived (a) SiCN1-1100. Reprinted with permission from

S. Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012).

Copyright 2012, American Chemical Society [136]; (b) SiCN2-1100 ceramics


4000 3500 3000 2500 2000 1500 1000 500

Inte

nsity

(a.

u.)

wavenumber (cm-1)

Figure 56 FT-IR spectrum of SiCN2-1100

At 1400 ºC, the 29Si MAS NMR spectrum of the SiCN1-1400 PDC has two main

bands representing SiC4 and SiN4 tetrahedra. The signal of SiC4 environments is located at ~ -

18 ppm. The formation of SiC phase is due to the carbothermal reaction of Si3N4 with carbon.

The embedded SiN4 nanodomains connect to “free” carbon domains via SiC4 domains, shown

in Figure 57 (a).

SiCN2-1400 shows an increase of the band at about -35 ppm in the 29Si MAS

spectrum, which represents SiN3C tetrahedra. With an increase in temperature, Si-N=C=N-

bonds break, nitrogen gas releases, and Si-C bonds form. 13C MAS NMR spectrum shows a

very broad band which is the overlap of the sp2 C-C and sp2 C-N bands. The schematic

representation of the bonding in SiCN2-1400 is shown in Figure 57 (b).

(a)

‐N=C=N‐


(b)

Figure 57 Schematic drawing of polysilylcarbodiimide-derived (a) SiCN1-1400, (b) SiCN2-1400 ceramics

4.4.1.2 Solid state structure of polyborosilylcarbodiimide-derived SiBCN1 and SiBCN2 ceramics

Besides 29Si MAS NMR (Figure 52 (c) and (d)) and 13C MAS NMR (Figure 53 (c) and

(d)), the structures of polyborosilylcarbodiimide-derived SiBCN1 and SiBCN2 ceramics were

also studied by 11B MAS NMR (Figure 54 (a) and (b)). Similar to SiCN1-1100 ceramic, the 29Si MAS NMR spectrum of SiBCN1-1100 (Figure 52 (c)) shows only one signal with a δiso

near -48 ppm indicating the presence of SiN4 tetrahedra, where silicon is bonded to four

nitrogen atoms as in crystalline Si3N4. This leads to the formation of Si3N4 domains that are

separated from the “free” carbon and boron nitride domains.

In SiBCN1 ceramics, sp2C-bonded carbon environments, characterized by a δiso of ~

125 ppm in the 13C MAS NMR spectra (Figure 53 (c)), take great part, but about 10% of the

carbon atoms are in form of C-N, which has a δiso of ~ 150 ppm. It is assumed that the C-N

bonds exist as the interface between the “free” carbon and the SiCxN4-x (x=0-4) domains in

SiBCN1 ceramics.

The 11B MAS spectrum of the SiBCN1 ceramics show only BN3 units characterized

by a δiso of ~ 27 ppm with a lack of other BCxN3-x (x=0-3) bonds. This means that BN

domains are located separately from the Si3N4 and C domains. Therefore, the structure of

SiBCN1 ceramics consists of Si3N4, BN and C nanodomains, which are connected through N–

B and N–C bonds. The BN domains most likely represent an interfacial phase between the

Si3N4 and C domains, as shown in Figure 58.


Figure 58 Schematic representation of SiBCN1-1100 and SiBCN2-1100 ceramics. From S. Widgeon, G. Mera,

Y. Gao, S. Sen, A. Navrotsky, and R. Riedel, J. Am. Ceram. Soc. 96, 1651 (2013). Copyright 2013 by John

Wiley Sons, Inc. Reprinted by permission of John Wiley & Sons, Inc [131].

For SiBCN2-1100, 29Si MAS NMR (Figure 52 (d)) shows only SiN4 units without

considerable Si-C bonds in the structure, while 13C MAS NMR (Figure 53 (d)) reveals another

signal with a δiso of ~ 150 ppm indicating C-N bonds besides the primarily sp2C-bonds. The

structure could be described to be the same as that of the SiBCN1 ceramic, as shown in Figure

58. Besides the similarity, SiBCN2 ceramic does show stronger intensity of the C-N band.

One possible reason for this could be because the SiBCN2 ceramic has much lower carbon

content, and thus a comparably higher amount of Si3N4 (see Chapter 3, Table 3 for the

chemical compositions of ceramics), and accordingly forms a higher volume of interfaces

with C-N bonds.

When the annealing temperature is increased to 1400 °C, both SiBCN1 and SiBCN2

ceramics show small bands in the 29Si NMR spectra at about -18 ppm suggesting the

formation of SiC domains. The signal is also increased at ca. 30 ppm in 13C MAS NMR

spectra which represents sp3 C-Si bonding. This was also seen by TEM and by selected area

electron diffraction (SAED) in previous studies of polyborosilylcarbodiimide-derived SiBCN

ceramics. The nanocrystalline SiC results from the solid-state reaction of amorphous C with

the Si3N4 domains (Section 4.2). Tavokoli et al. have observed the same crystallization

behavior [152].

4.4.1.3 Solid state structure of polysilazane-derived SiCN3 and SiCN4 ceramics

Polysilazane-derived SiCN3 and SiCN4 ceramics prepared at 1100°C have mixed

bonding environments about the Si atoms, namely SiCxN4-x (x=0~4) tetrahedra. This is in

sharp contrast to the polysilylcarbodiimide-derived SiCN ceramics which have no such mixed

bonding environments. In the 29Si MAS NMR spectra, the chemical shifts of the SiCN3,

SiC2N2, SiC3N tetrahedra are centered at ca. -35, -5, and 3 ppm, respectively. The 29Si MAS

NMR spectra of SiCN3 and SiCN4 ceramics prepared at 1100°C (Figure 52 (e) and (f)) show


an overlap of several bands which are from SiCxN4-x (x=0-4) mixed bonds. The band at -18

ppm, characteristic of the SiC4 tetrahedra, is also observed in the structure of samples

pyrolyzed at 1100°C. The 13C MAS NMR spectra of these two ceramics (Figure 53 (e) and (f))

show primarily two resonances at ca. 125 and 150 ppm indicating sp2C-C environments, as

well as sp2C-N bonds. The 13C MAS NMR spectra also confirm the presence of sp3C-Si

bonds which have a δiso of ~ 30 ppm. The presence of SiC4 tetrahedra indicates the formation

of SiC nanodomains at 1100°C. This is different from polysilylcarbodiimide-derived SiCN1

and SiCN2 ceramics, which contain no SiC nanodomains after pyrolysis at the same

temperature. Figure 59 presents a schematic of the structure of SiCN3 and SiCN4 ceramics

prepared at 1100°C. The SiC nanodomain is assumed to be the transition region between the

Si3N4 and carbon nanodomains.

Figure 59 Schematic representation of SiCN3-1100 and SiCN4-1100 ceramics

In the 29Si MAS NMR spectra of ceramics prepared at 1400°C, the intensity of the

band at -18 ppm, which represents the SiC4 tetrahedral environments, increases relative to the

spectra of samples pyrolyzed at lower temperatures. This demonstrates the growth of SiC with

an increase of annealing temperature.

4.4.1.4 Solid state structure of polyborosilazane-derived SiBCN3 and SiBCN4 ceramics

The 29Si MAS NMR spectra of polyborosilazane-derived SiBCN3 and SiBCN4

ceramics (Figure 52 (g) and (h)) reveal the presence of significant mixed bonded SiCxN4-x

tetrahedra, similar to in the boron-free analogues. SiBCN3 and SiBCN4 ceramics also show

BCxN3-x mixed bonded environments (Figure 54 (c) and (d)). In comparison to the SiBCN3

ceramics, the SiBCN4 ceramic prepared at 1100ºC contains a higher relative fraction of

carbon-rich SiCxN4-x (x ≥2) tetrahedral units, although this sample actually has a lower carbon

content. This observation may be explained by the condensation reaction among methyl

groups to form Si–C–Si bonds in the precursor. This condensation does not occur in the

SiBCN3-1100 sample because the substitute is a phenyl group in the precursor.


The 29Si CPMAS spectrum of the SiBCN4-1100 sample (Figure 60) shows that these

C-rich SiCxN4-x tetrahedral units are preferentially protonated with respect to the N-rich

tetrahedra. Therefore, these two types of tetrahedra must be spatially segregated in the

structure. It is assumed that SiCxN4-x (x=0-4) domains have cores composed primarily of SiN4

tetrahedra, while the carbon-rich SiCxN4-x units are located near the boundary of these

domains. These units may be directly connected to the carbon domains via Si-C-C bonds, or

alternatively via BN, which may make up the interface between the SiCxN4-x (x=0-4) and

“free” carbon domains. In such a BN transition zone, BN3 units are located near the core of

this interface region. The BN transition zone is bonded to SiCxN4-x (x=0-4) domains via Si–

C–B linkages, and to “free” carbon domains via B-C-C or N-C-C bonds. However, there is a

lack of mixed BCxN3-x units, except for BCN2 unit, which precludes the possibility of a

homogeneous intermixing of B, N, and C atoms in graphene/borazine-like sheets. The 13C

MAS NMR spectra (Figure 53 (g) and (h)) show a sp2 C network with significant amounts of

sp2 C–N and sp2 C–B bonding. This is completely consistent with the structural model

proposed above, for which a schematic is shown in Figure 61. Finally, the SiBCN4-1100

sample has relatively higher concentrations of interfacial sp2 C–N and sp2 C–B environments,

as well as of BCN2 units, compared to the SiBCN3-1100 sample. This observation suggests

that low carbon concentrations lead to the formation of smaller carbon domains. Smaller

carbon domains result in a higher volume fraction of interfaces among the domains, where sp2

C–N and sp2 C–B environments are present (also seen in Scarlett Widgeon et al. 2013).

40 20 0 -20 -40 -60 -80 -100 -120


Figure 60 29Si CPMAS NMR (solid line) and 29Si MAS NMR (dash line) of SiBCN4-1100 ceramic. From S.

Widgeon, G. Mera, Y. Gao, S. Sen, A. Navrotsky, and R. Riedel, J. Am. Ceram. Soc. 96, 1651 (2013).

Copyright 2013 by John Wiley Sons, Inc. Reprinted by permission of John Wiley & Sons, Inc [131].


Figure 61 Schematic representation of SiBCN3-1100 and SiBCN4-1100 ceramics. From S. Widgeon, G. Mera,

Y. Gao, S. Sen, A. Navrotsky, and R. Riedel, J. Am. Ceram. Soc. 96, 1651 (2013). Copyright 2013 by John

Wiley Sons, Inc. Reprinted by permission of John Wiley & Sons, Inc [131].

In SiBCN4-1400 samples, the relative amount of interfacial sp2 C–N and sp2 C–B

bonds decreases, while that of B-N and C-C bonds increases. The intensities of the bands,

which represent SiCxN4-x (x=0-4) and BCxN3-x (x=0-3) mixed bonding environments,

decrease. This is to say, the ceramics undergo a demixing process and a nanodomain

coarsening. Details are discussed in Section 4.3. Calorimetry demonstrates that this structural

evolution is favorable in both enthalpy and free energy.

4.4.2 Microstructure of Si(B)CN ceramics

In small angle X-ray scattering (SAXS), X-rays are used to investigate the structural

properties of materials. Photons interact with electrons, and provide information about the

fluctuations of electronic densities in heterogeneous matter. In the SAXS of PDCs, the

intensity of the scattering signal is propotional to the electronic contrast between the domains.

In this way, the nanodomains of PDCs can be detected using this technique. By studying the

change in scattering intensity, at least two structural levels in all ceramic samples prepared at

1100 ºC and 1400 ºC in this study were confirmed (Figure 62). Unfortunately, we were not

able to collect data over the entire q range relevant for these materials. Data is lacking at low

q, where the Guinier region of the first structural level (larger domain size) would be found,

as well as at high q, where the Porod region for the second structural level (smaller domain


size) would be found. However, relatively complete data is available for the Guinier region of

the second structural level, and a fitting was performed in this region.

1

SiBCN1-1400

SiBCN1-1100

q (nm-1)

SiCN1-1100

SiCN1-1400

1

SiCN2-1100

SiBCN2-1400

SiCN2-1400

q (nm-1)

SiBCN2-1100

1

SiBCN3-1100

SiCN3-1100

SiCN3-1400

q (nm-1)

SiBCN3-1400

1

SiBCN4-1100SiBCN4-1400

SiCN4-1100

q (nm-1)

SiCN4-1400

Figure 62 SAXS results of ceramics prepared at 1100°C and 1400°C

The Beaucage equation (Formula 4) was used for data fitting (Figure 63) [169]. The

variables G, R, B and P were fitted using Origin 8.1 for the second structural level. Here R

represents Rg, which is the radius of gyration from the domains in the second structural level.

.

Figure 63 Beaucage equation for fitting SAXS data

Second structural levelFirst structural level


≅ √

Formula 4

The radius of gyration (Rg) values obtained from fitting the second structural level of

the SAXS data are shown in Table 11. The Rg values of ceramics prepared at 1100C are

mostly 0.5 - 1.5 nm, while the Rg values of the ceramics prepared at 1400C mostly range

from 1.5 to 3 nm, indicating domain coarsening at higher pyrolysis temperatures.

Table 11 Radius of gyration of ceramics prepared at 1100C and 1400C

Ceramic sample Formula of preceramic

precursor

Rg(nm) of the ceramics

prepared at 1100C

Rg(nm) of the ceramics

prepared at 1400C

SiCN1 [-(PhVi)Si-NCN-]n 0.69 1.26

SiBCN1 [-(MeVi)Si-NCN-]n 0.45 1.19

SiCN2 B[-(C2H4)Si(Ph)-NCN-]3n 1.21 2.48

SiBCN2 B[-(C2H4)Si(Me)-NCN-]3n 1.02 1.87

SiCN3 [-(PhVi)Si-NH-]n 1.37 2.98

SiBCN3 [-(MeVi)Si-NH-]n 1.28 2.79

SiCN4 B[-(C2H4)Si(Ph)-NH-]3n 2.17 3.76

SiBCN4 B[-(C2H4)Si(Me)-NH-]3n 1.21 2.28

Based on previous studies, nanodomains in carbon rich polymer-derived SiCN

ceramics normally consist of SiCxN4-x (x=0-4) domains and “free” carbon domains. SiBCN

ceramics also have nanosized features containing B, N and C. The carbon rich ceramic

samples in this work contain a large amount of “free” carbon, varying from 40 mol % to 90

mol % (Table 3). TEM observations (SiCN1-1100 and SiCN1-1400 ceramics as an example,

shown in Figure 64) illustrate that the structure of these samples is amorphous, though at

1400C, the ordering of few graphene layers is observed (Figure 64). Numerous TEM

observations indicate that the stacking number of the graphene sheets is ~ 1-5 in ceramics

pyrolyzed at 1100 and 1400ºC.

Figure 64 TEM observation of carbon rich SiCN1 PDCs prepared at (a) 1100C and (b) 1400C

5nm 5nm


One parameter used to characterize the extent of graphitization is “La,” which can be

obtained from Raman spectroscopy. La is the average lateral size of the graphitic carbon. The

formula for determining La and additional explanations are given in Section 4.2. The Raman

spectra of ceramic samples prepared at both 1100 and 1400C are shown in Figure 65. The

calculated La values are also shown in the same figure. The average lateral size along the six-

fold rings of graphitic carbon is in the range of 1.4-2.2 nm. It is notineable that La values of

ceramics pyrolyzed at 1100 and 1400 ºC do not show the same trend as the Rg values

determined by SAXS. Rg of ceramics pyrolyzed at 1400 ºC are always larger than those of

ceramics pyrolyzed at 1100 ºC.

500 1000 1500 2000 2500 3000 3500 4000

SiBCN1-1400

SiCN1-1400

SiBCN1-1100

SiCN1-1100La = 2.15 nm

La = 1.41 nm

La = 1.87 nm

Raman shift (cm-1)

Inte

nsi

ty(a

.u.)

La = 2.18 nm

500 1000 1500 2000 2500 3000 3500 4000

SiBCN2-1400

SiCN2-1400

SiBCN2-1100

SiCN2-1100

La = 1.44 nm

La = 1.62 nm

La = 2.20 nm

Inte

nsi

ty(a

.u.)

Raman shift (cm-1)

La = 1.23 nm

500 1000 1500 2000 2500 3000 3500 4000

SiBCN3-1400

SiCN3-1400

SiBCN3-1100

SiCN3-1100La = 1.89 nm

La = 2.10 nm

La = 1.82 nm

La = 1.90 nm

Raman shift (cm-1)

Inte

nsi

ty (

a.u

.)

500 1000 1500 2000 2500 3000 3500 4000

SiBCN4-1400

SiCN4-1400

SiBCN4-1100

SiCN4-1100La = 1.85 nm

La = 1.74 nm

La = 2.11 nm

La = 1.89 nm

Inte

nsi

ty (

a.u

.)

Raman shift (cm-1) Figure 65 Raman spectroscopy of (a) SiCN1, SiBCN1, (b) SiCN2, SiBCN2, (c) SiCN3, SiBCN3, (d) SiCN4,

SiBCN4 PDCs

In fact, La is the lateral size of graphene layers without tortuosities. La does not

characterize the full length of the graphitic carbon if the sheets are tortuous [170]. It is

observed that the graphitic carbon in carbon-rich amorphous polymer-derived ceramics is

always tortuous. Thus, La cannot be used to characterize the carbon domain sizes in these

materials. Tortuous graphene sheets with stacking numbers from 1 to 5 form a network in the

polymer-derived ceramic structure. The SiCxN4-x (x=0-4) domains are located among the

network of carbon sheets. A schematic of this nanodomain feature is shown in Figure 66.

This schematic is also consistent with a previously published model [161].

(a) (b)

(c) (d)


Figure 66 Schematic of nanodomain structure of carbon rich PDCs

Provided that the SiCxN4-x (x=0-4) domains are spherical, the size of these

nanodomains can be calculated from Rg using the equation: R R . The

distribution of nanodomain sizes calculated in this way are shown in Figure 67.

1100 1200 1300 1400

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

SiCN1SiBCN1

SiBCN2

SiBCN4SiCN2SiBCN3SiCN3

Temperature (C)

Rsp

her

e(n

m)

SiCN4

Figure 67 Si nanodomains size in ceramics prepared at 1100C and 1400C

As is well-known, Si(B)CN ceramics prepared at 1100 °C typically remain amorphous.

This is certainly true for the materials invested in this work, as demonstrated by XRD in

Figure 68. Meanwhile, some of the ceramics pyrolyzed at 1400°C show nanocrystallization,

as is evidenced by broad humps in the XRD pattern. The reflections at 2θ = 37º, 41º, 60º, 72º

are from the (111), (200), (220), and (311) planes of SiC. The average size of the SiC crystals

can be estimated using the Scherrer equation. Taking SiBCN3-1400 and SiCN4-1400 as two

examples, the sizes of the SiC nanocrystals in SiBCN3-1400 and SiCN4-1400 ceramics are ca.

2.4 nm and 2.8 nm, respectively, if no other contributions to peak broadening are considered.

The Si domain sizes obtained by SAXS of these two ceramics are 3.6 nm and 4.8 nm, which


is bigger than the size of SiC nanocrystals. These results suggest that the nanodomain sizes

obtained by SAXS are representative of the SiCxN4-x (x=0-4) nanodomains, which should be

larger than the crystallites.

10 20 30 40 50 60 70 80

SiCN1-1400

SiBCN1-1400

SiBCN1-1100Inte

nsi

ty (

a.u

.)

2

SiCN1-1100

10 20 30 40 50 60 70 80

Inte

nsi

ty (

a.u

.)

2

SiCN2-1400

SiBCN2-1400

SiBCN2-1100

SiCN2-1100

10 20 30 40 50 60 70 80

Inte

nsi

ty(a

.u.)

2

SiCN3-1400SiBCN3-1400

SiBCN3-1100

SiCN3-1100

10 20 30 40 50 60 70 80

Inte

ns

ity

(a.u

.)

2

SiCN4-1400

SiBCN4-1400

SiBCN4-1100SiCN4-1100

Figure 68 XRD patterns of (a) SiCN1, SiBCN1, (b) SiCN2, SiBCN2, (c) SiCN3, SiBCN3, (d) SiCN4, SiBCN4

It was also found that the average sizes of the SiCxN4-x (x=0-4) domains (Figure 67) in

ceramic samples derived from different precursors show some interesting characteristic:

First, the SiCxN4-x (x=0-4) domains in all ceramic samples pyrolyzed at 1400 C are

larger than in the analogues pyrolyzed at 1100C. This coarsening phenomenon with

increasing temperature has already been studied and reported. SiCxN4-x (x=0-4) domains grow

with the concomitant elimination of interfaces.

Second, boron modified ceramics have smaller SiCxN4-x (x=0-4) domains than that of

the boron-free ceramics. Generally after thermolysis, boron forms BN, so fewer nitrogen

atoms connect to silicon atoms to form SiN4 tetrahedra in the SiBCN samples than in their

SiCN counterparts. At the macro scale, the amount of Si-N4 is reduced. BN is considered to

be located at the interface between Si and “free” carbon domains. The electron density of the

BN unit is similar to that of carbon, so there is no contrast under X-ray scattering. BN is

considered to be in the same region as that of the carbon domains. The reduced amount of

SiN4 tetrahedra may result in smaller SiCxN4-x (x=0-4) domains, but not necessarily. Another

explanation for the smaller SiCxN4-x (x=0-4) domains can be traced back to the synthesis of

(a) (b)

(c) (d)


the boron-modified polymers. During hydroboration, boron atoms connect to vinyl groups. In

this way, vinyl groups do not connect to each other, as they do during the polymerization of

boron-free polymers. Therefore, in the boron-modified polymers, fewer Si atoms gather via

the polymerization process, but are separated by boron atoms. These boron atoms form BN

during pyrolysis. This BN “barrier” leads to smaller SiCxN4-x (x=0-4) domains in the ceramics.

A third trend revealed by SAXS is that if the preceramic precursors contain the same

organic substitute attached to Si, the SiCxN4-x (x = 0-4) domains in the ceramics derived from

poly(boro)silylcarbodiimides are always smaller than those in the ceramics derived from

poly(boro)silazanes. One explanation for this trend is that, if the polymer is crosslinked at

100 to 300 ºC, nitrogen atoms of the Si-N(H)-Si groups in poly(boro)silazanes can connect to

a third silicon atom through transamination reaction. In any case, N-H is involved in forming

the network structure. In the case of poly(boro)silylcarbodiimide, the decomposition of -

N=C=N- groups can still be observed even up to 950 ºC with the evolution of HCN (e.g. in [-

(PhVi)Si-NCN-]n polymer as shown in Figure 69). The breaking of double bonds in N=C=N

groups separates SiCxN4-x (x=0-4) domains from “free” carbon domains. This separation is

also shown in the 29Si MAS NMR spectra of the SiCN1 ceramics (Figure 52 (a)). A schematic

is shown in Figure 55(a) [136]. The activity of the N=C=N groups causes the final SiCxN4-x

(x=0-4) domains in poly(boro)silylcarbodiimides-derived ceramics to be smaller.

200 400 600 800 1000 1200 1400

Inte

nsi

ty (

a.u

.)

Temperature (C)

Figure 69 m/z=27 fragment in Mass Spectra of [-(PhVi)Si-NCN-]n polymer

At last, the carbon content also affects the SiCxN4-x (x=0-4) domain size in the

ceramics. Most obviously, samples with lower carbon contents tend to have larger SiCxN4-x

(x=0-4) domains, except the case of the SiBCN3 and SiBCN4 ceramics. Differences in carbon

content arise due to the use of different substitutes in the purecursor polymers (i.e. phenyl

group and methyl group). During the crosslinking process and the following pyrolysis, phenyl

groups sterically hinder the connection between silicon atoms and the formation of Si

networks. In other words, Si domains are restricted to relatively small sizes because phenyl

groups are too large to allow Si tetrahedral units to come into close contact. However, in

methyl-substituted polymers, it is easier for Si tetrahedral units to interact to form larger


SiCxN4-x (x=0-4) domains. The opposite trend was found in case of the SiBCN3 and SiBCN4

ceramics. SiBCN4-1100, which is derived from methyl-substituted polyborosilazane, has

SiCxN4-x (x=0-4) domains similar in size to those of SiBCN3-1100, which is derived from

phenyl-substituted polyborosilazane. SiBCN4-1400 has smaller SiCxN4-x (x=0-4) domains

than SiBCN3-1400. The reason for this remains unclear.

4.4.3 Summary

The nano-scale structure of polymer-derived ceramics strongly depends on the nature

of their precursor polymers. By using a suite of characterization techniques, information

regarding domain composition, domain size, and interdomain bonding could be obtained.

MAS NMR spectroscopic investigations support the finding that ceramics derived from

polysilazanes contain silicon mixed bonding environments, namely SiCxN4-x(x=0-4)

tetrahedra, while ceramics derived from polysilylcarbodiimides have no such mixed bonding.

The boron-modified counterparts show the same phenomenon. While all of the synthesized

ceramics are basically made up of SiN4 and sp2 carbon nanodomains, the nature of the

transition zones between these domains differ depending on the architecture of the precursor.

Polysilylcarbodiimide-derived SiCN1 and SiCN2 ceramics pyrolyzed at 1100C contain

Si3N4 and “free” carbon domains. The SiCN1 ceramic consists of a continuous carbon phase

in which Si3N4 domains are embedded, but no considerable amount of chemical bonding

exists between these domains. In contrast, in the SiCN2 ceramics, these domains are

connected by C-N bonds. Boron-modified polyborosilylcarbodiimide-derived SiBCN1 and

SiBCN2 ceramics prepared at 1100C contain Si3N4 and “free” carbon domains, as well as an

additional BN phase which exists at the interface between these two domains. Consequently,

the Si3N4 and “free” carbon domains connect to the BN phase via N-B and N-C bonds. For

polysilazane-derived SiCN3 and SiCN4 ceramics pyrolyzed at 1100C, mixed SiCxN4-

x(x=1~4) tetrahedra are found located interfacially between the SiN4 and “free” carbon

domains. SiC4 tetrahedra are connected to the “free” carbon domains. The MAS NMR spectra

of the polyborosilazane-derived SiBCN3 and SiBCN4 ceramics prepared at 1100C indicate

that there are SiCxN4-x(x=0~4) tetrahedra and BCxN3-x (x=0~4) mixed bonds which connect

the Si3N4 domains to the “free” carbon domains via C-B bonds (between the SiC4 tetrahedron

and the BN phase) and B-C bonds (between the BN phase and the “free” carbon domain).

Considering results from Raman spectroscopy, XRD, and TEM, the radius of gyration

values obtained by SAXS are considered to be the size of the SiCxN4-x (x=0-4) nanodomains

in the amorphous PDCs. Trends in the domain size with respect to the type of precursor are

summarized as follows: 1) for the materials studied, Si-containing domain sizes increase with

an increase in the pyrolysis temperature from 1100 ºC to 1400 ºC, indicative of coarsening

during heat treatment; 2) boron-modified ceramics have smaller Si-containing domains than

that of the boron-free ceramics; 3) the SiCxN4-x (x=0-4) domains in the ceramics derived from

poly(boro)silylcarbodiimides are always smaller than those derived from poly(boro)silazanes;


4) lower carbon contents result in larger SiCxN4-x (x=0-4) domains, except for

polyborosilazane-derived SiBCN ceramics.

Electrical / electrochemical properties 105

4.5 Electrical / electrochemical properties

4.5.1 Complex impedance spectra of selected polymer-derived ceramics

Complex impedance spectroscopy (IS) is a useful tool for studying the electrical

behavior of composite materials [171]. Unlike DC resistance measurements, which study the

average behavior of a sample, impedance spectroscopy is sensitive to phase differences. The

parameters derived from IS, such as conductance and capacitance, can often be correlated to

complex materials variables such as composition, microstructure, mass transport, etc. In

general, each phase of a composite can be represented by an equivalent electrical sub-circuit

containing a resistor and a capacitor; and the impedance of the entire composite can then be

represented by a unique equivalent circuit comprised of these sub-circuits.

Polymer-derived ceramics can be considered multiphase composite materials. Their

electric conduction behavior strongly depends on the distribution and conductivity of each

phase. While the electrical behavior of polymer-derived ceramics has been studied to a certain

extent, previous studies primarily focused on measuring the DC resistance of the entire

material. This type of measurement has two main drawbacks: (i) it cannot determine the

contribution of each phase to the total conductivity of the material, and (ii) it cannot be used

to determine the structure based conduction mechanisms.

In this section, we study the impedance behavior of selected SiCN PDCs. If we

consider a polymer-derived ceramic as a two-phase composite [172], then the possible paths

for electric current depend on whether the two phases are continuous. If both phases are not

continuous, as shown in Figure 70 (a1), electric current within the material will have only one

possible path via both phases in series (Figure 70 (b1)). If one phase is continuous and the

other is not, as shown in Figure 70 (a2), electric current has two possible paths as shown in

Figure 70 (b2): conduction via the first phase (the continuous one), or conduction via the first

and the second phase in series. If both phases are continuous, electric current within the

material will have three possible paths as shown in Figure 70 (a3): conduction via the first

phase; conduction via the first and the second phase in series, and conduction via the second

phase. The corresponding equivalent circuit is shown in Figure 70 (b3). The IS results will be

analyzed using this model.

(a1) (b1)

106 Electrical / electrochemical properties

(a2) (b2)

(a3) (b3)

Figure 70 (a) Schematic models of two-phase composite with conduction paths: (a1) both phases are not continuous; (a2) one phase is continuous, the other is not; (a3) both phases are continuous; and (b) the corresponding equivalent electrical circuits of (a1), (a2) and (a3).

Three polymer-derived ceramics have been selected as model materials for studying

impedance behavior: SiCN4, SiBCN2 and SiBCN4. The compositions of the three materials

prepared at 1100 and 1400C are listed in Table 12. It is seen that the major conducting

phases in the studied ceramics are SiC and “free” carbon, because Si3N4, SiO2 and BN are

considered insulating phases. SiBCN2 and SiBCN4, in comparison to SiCN4, contain a fairly

large amount of BN, while SiCN4 contains no BN phase. Besides, SiBCN4 consists mainly of

“free” carbon and a SiC phase, while the major phase in SiBCN2 is “free” carbon with a small

amount of SiC. As discussed in section 4.4.1.3 and 4.4.1.4, poly(boro)silazane derived

Si(B)CN ceramics have domains consisting of mixed bonded SiCxN4-x(x=0-4) tetrahedra.

These SiCxN4-x (x=0-4) domains have cores composed primarily of SiN4 tetrahedra, while the

SiC4 tetrahedra are located at the boundary of these domains. In SiCN4 and SiBCN4 ceramics,

a considerable amount of SiC4 tetrahedra are present, which is revealed by the intense band

centered at -18 ppm in the 29Si MAS NMR spectra (Figure 52 (f) and (h)). At 1400°C, all

three studied ceramic samples contain SiC nanocrystallites, as evidenced by XRD (SiBCN2 in

Figure 68 (b), SiCN4 and SiBCN4 in Figure 68 (d)).


Table 12 Compositions of the samples investigated with for impedance spectroscopy

This table is a part of Table 3.

† Samples have not been measured on boron element. The values are estimated from the ratio of starters in polymer reactions and the ceramic yield.


Si wt%

atom%

C wt%

atom%

N wt%

atom%

B wt%

atom%

O wt%

atom%

H wt%

atom%

Cl wt%



Free C

Si3N4 SiC BN SiO2

B[-(C2H4)Si(Me)-NCN-]3n

SiBCN2-1100 37.71 21.78

30.41 40.98

24.67 28.50

3.00† 4.49†

4.21 4.25

— — SiC1.88N1.31B0.21O0.20 73.38 11.20 3.08 8.38 3.97

SiBCN2-1400 39.94 23.34

30.31 41.33

23.43 27.39

3.00† 4.55†

3.32 3.40

— — SiC1.77N1.17B0.19O0.15 69.10 10.72 8.47 8.53 3.19

[-(MeVi)Si-NH-]n

SiCN4-1100 51.75 31.01

29.19 40.81

17.10 20.49

— 1.56 1.64

0.36 6.04

0.04 0.02

SiC1.32N0.66O0.05 55.6 11.0 31.7 — 1.8

SiCN4-1400 51.283 31.55

29.84 42.84

17.23 21.20

— 1.44 1.55

0.165 2.84

0.042 0.02

SiC1.36N0.67O0.05 57.2 10.8 30.4 — 1.6


SiBCN4-1100 48.36 29.70

28.85 41.35

13.60 16.71

4.52 7.07

4.63 4.98

0.01 0.17

0.03 0.01

SiC1.39N0.56B0.24O0.17 40.1 4.5 37.5 13.3 4.7

SiBCN4-1400 48.20 29.55

28.39 40.62

14.09 17.28

4.68 7.44

4.59 4.93

0.01 0.17

0.04 0.02

SiC1.37N0.58B0.25O0.17 39.5 4.6 37.2 14.0 4.7


Figure 71, Figure 72, Figure 73 show the impedance spectra of SiCN4, SiBCN2 and

SiBCN4 materials, respectively, as organized by pyrolysis temperature. The complex

impedance plots typically take the form of semicircular arcs. In some cases, the arcs are

incomplete in the high frequency region due to limitations of the experimental facilities. The

spectrum for SiCN4-1100 exhibits one big semicircle, while the spectra for SiCN4 samples

prepared at 1200°C, 1300°C and 1400°C exhibit one semicircle at low frequency followed by

an arc obtained at higher frequency. This arc represents the beginning of another semicircle.

The two semicircles are partially overlapping. The spectra of samples containing a BN phase,

namely SiBCN2 and SiBCN4, show only one semicircle or an incomplete semicircular arc.

0 1x105 2x105 3x105 4x105

0

1x105

2x105

3x105

4x105

‐Z"(Ohm)

Z'(Ohm)

SiCN4-1100

0 500 1000 1500 2000 2500 3000

0

500

1000

1500

2000

‐Z"(Ohm)

Z'(Ohm)

SiCN4-1200

0 100 200 300 400 500 600 700

0

100

200

300

400

500

600

Z'(Ohm)

‐Z"(Ohm)

SiCN4-1300

0 100 200 300 400 500 600

0

50

100

150

200

250

300

350

400

Z'(Ohm)

‐Z"(Ohm)

SiCN4-1400

Figure 71 Complex impedance spectra of SiCN4 samples pyrolyzed at different temperatures.

0 500 1000 1500 2000 2500 3000 3500

0

500

1000

1500

2000

2500

3000

Z'(Ohm)

‐Z"(Ohm)

SiBCN2-1100

0 500 1000 1500 2000 2500

0

500

1000

1500

2000

‐Z"(Ohm)

Z'(Ohm)

SiBCN2-1200

ω ω

ω

ω

ω ω


0 100 200 300 400 500 600

0

100

200

300

400

500‐Z

"(Ohm)

Z'(Ohm)

SiBCN2-1300

0 50 100 150 200

0

20

40

60

80

100

120

140

‐Z"(Ohm)

Z'(Ohm)

SiBCN2-1400

Figure 72 Complex impedance spectra of SiBCN2 samples pyrolyzed at different temperatures.

0 500 1000 1500 2000 2500 3000

0

500

1000

1500

2000

2500

Z'(Ohms)

‐Z"(Ohms)

SiBCN4-1100

0 50 100 150 200 250 300

0

50

100

150

Z'(Ohms)

‐Z"(Ohms)

SiBCN4-1200

0 50 100 150 200 250 300

0

50

100

150

200

Z'(Ohms)

‐Z"(Ohms)

SiBCN4-1300

0 10 20 30 40 50

0

5

10

15

20

25

Z'(Ohms)

‐Z"(Ohms)

SiBCN4-1400

Figure 73 Complex impedance spectra of SiBCN4 samples pyrolyzed at different temperatures.

The most accepted approach to interpret the semicircle phenomena is to consider them

with respect to the distribution of relaxation time. The semicircles measured for the

investigated materials are modeled by an electrical equivalence consisting of a parallel

resistance and capacitance circuit corresponding to the individual component of the material.

The parameters which were used to obtain the best fit for the experimental results are listed in

Table 13. It is found that the two-semicircle spectra obtained for SiCN4 samples can be best

fitted by the equivalent circuit shown in Figure 74(a), which contains two resistor/capacitor

sub-circuits in series. The same electrical equivalence is used to model the spectrum of

ω

ω

ω

ω

ω ω


SiCN4-1100 sample, although its spectrum resembles one semicircle. However, analying this

spectrum with only one resistor/capacitor circuit resulted in poorer fitting. The goodness of

the fit is given by the parameters, “Chi-squared” and “sum of sqr”, as listed for both cases in

Table 13. The equivalent circuit with two resistor/capacitor sub-circuits in series yields lower

values for “Chi-squared” and “sum of sqr” (i.e. greater goodness of fit) than that of the

equivalent circuit with one resistor/capacitor. The presence of two semicircles is also

consistent with the spectra measured on the same ceramic prepared at other temperatures. The

two semicircle phenomenon is similar to those previously measured for polymer-derived

silicon oxycarbide (SiOC) ceramics [172], suggesting the presence of two phases. The

semicircle measured at low frequency is considered to represent conduction in the “free”

carbon phase, while the semicircle at high frequency is attributed to the Si-containing phase.

The spectra for SiBCN2 and SiBCN4 samples prepared at all studied temperatures can

be estimated and best fitted as one semicircle. This semicircle represents the equivalent circuit

shown in Figure 74(b), which contains one resistor/capacitor sub-circuit. Therefore, we

assume that the conduction of these two samples is dominated by one phase. For some

samples, suppressed semicircles are represented in the Nyquist plots. This suggests that the

studied materials are not ideal capacitors, and a constant phase element (CPE) is required to

obtain the best fitting (Figure 74). The suppressed semicircles have been explained by a

number of phenomena, depending on the nature of the system being investigated. However, a

common thread among these explanations is that some property of the system is not

homogeneous, or that there is some distribution (dispersion) of the value of a physical

property of the system. CPE consists of a capacitive element CPE-T and of CPE-P (0-1),

which is a measure of how much the CPE deviates from an ideal capacitor. When CPE-P = 1,

the CPE is an ideal capacitor and can be replaced by C.

Table 13 Properties of elements from the equivalent circuits*

Sample Elements 1100C 1200C 1300C 1400C

SiCN4

R1(Ω) 17137 826.8 337.3 203.4

C1/CPE1(F) CPE‐T 1.888 × 10‐10

CPE‐P 0.90799 1.396 × 10

‐11 1.664 × 10‐11 1.822× 10‐11

R2 (Ω) 391640 1879 374.5 378.1

C2/CPE2(F) CPE‐T 2.227 × 10‐11

CPE‐P 0.96435 4.175 × 10‐10 5.155 × 10‐10 4.335 × 10‐10

Chi‐Squared /Sum of Sqr

0.0011413 /0.1187

0.0010251 /0.057408

0.00022046 /0.012346

0.00010317 /0.0057776

Chi-Squared /Sum of Sqr (For one RC

fit)

0.015858 /1.6968 — — —

SiBCN2

R (Ω) 3358 2314 592.7 236

C/CPE(F) 1.159 × 10‐11 5.265 × 10‐10 1.044 × 10‐11

CPE‐T 1.398 × 10‐10

CPE‐P 0.88652


0.0013527 /0.02976

0.0013994 /0.13714

0.0000564 /0.00090315

0.0000874 /0.0020118


Sample Elements 1100C 1200C 1300C 1400C

SiBCN4

R (Ω) 2804 271.3 264.7 42.48

C/CPE(F) 2.299 × 10‐11 1.446 × 10‐11 2.164 × 10‐11

CPE‐T 1.249 × 10‐8

CPE‐P 0.68394


0.0034681 /0.20115

0.0026268 /0.15235

0.00025252 /0.014646

0.00024328 /0.0080283

* All the values were obtained base on the normalized size of samples: diameter 8mm, thickness 0.5mm.

(a)

(b)

Figure 74 equivalent circuits used to fit the experimental impedance spectra.

As discussed above, the major conductive phases in the studied ceramics are SiC and

“free” carbon. The impedance measurements confirm that the resistance of the ceramic

decreases with increasing pyrolysis temperature. It is explained by the increased ordering of

both the “free” carbon and SiC phases with increasing pyrolysis temperature, which leads to

higher conductivity.In SiBCN2 samples, the major phase is “free” carbon, so it is likely the

phase which dominates the conduction of SiBCN2. SiBCN4 samples have a lower “free”

carbon concentration, but the overall conductance is higher than that of SiBCN2 (Figure 75).

Noticing that the SiBCN4 samples contain a relatively large amount of the SiC phase, the SiC

phase is considered to be responsible for their higher conductivity. This means that SiC is the

dominant conductive phase in the SiBCN4 samples. The SiC dominant position is also

indicated by the strong increase in conductivity after pyrolysis at 1400°C, when the sample

contains nano crystalline SiC. This one-phase dominant conduction phenomenon is not found

in SiCN4 samples which contain similar amount of “free” carbon and SiC, as in the SiBCN4

samples.


1100 1200 1300 14001E-3

0,01

0,1

SiBCN2 SiBCN4

Temperature (C)

(S/m

)

Figure 75 Comparison of overall conductance of SiBCN2 and SiBCN4 as a function of pyrolysis temperature.

To understand the different conduction mechanism in these samples, nanodomain

structure is considered. As discussed in Section 4.4, NMR studies revealed that the Si-

containing phase exhibits a microstructure wherein the Si3N4 forms a core and SiC forms a

shell about it. For SiBCN samples, the BN phase is located between the Si-containing and the

“free” carbon phases. Based on the structural information, we propose the following simple

structural models to account for the observed conduction behavior of these materials (Figure

76). In SiCN4 (Figure 76 (a)), neither SiC nor the carbon phase forms a continuous network.

Therefore, the conduction path goes via both phases in series. In SiBCN2 (Figure 76 (b)), the

“free” carbon forms a continuous phase, and the BN forms at the interfacial region between

the matrix “free” carbon and the isolated SiC phases. The BN phase isolates the SiC phase

from the “free” carbon phase, thus the conduction path proceeds via the “free” carbon phase.

In SiBCN2-1400 sample, even though the amount of SiC increases and SiC shows slight

nanocrystallinity, the proposed mechanism can still be applied. In SiBCN4 (Figure 76 (c)), the

SiC forms a continuous phase, and the BN forms the interfacial region between the isolated

“free” carbon and the matrix SiC phases. Again, the BN phase isolates the “free” carbon

phase from the SiC phase, and the conduction path goes via the SiC phase. The higher

conductivity of SiBCN4 as compared to SiBCN2 is likely because the conductivity of the SiC

phase is higher than that of the “free” carbon phase. This can also be seen from the SiCN4

samples: the resistance of the SiC phase (R1) is lower than that of the “free” carbon phase

(R2), even though the concentrations of the two phases are similar.


Figure 76 Schematic model showing the microstructures for (a) SiCN4, (b) SiBCN2 and (c) SiBCN4. “Other phases” represent the non-conductive Si3N4, SiO2, etc.

In summary, complex impedance spectroscopy was used to study the electrical

behavior of three polymer-derived ceramics synthesized in this work. The experimental data

were analyzed by modeling the processes by means of electrical equivalence circuits. By

doing so, microstructure-based conduction mechanisma are revealed. In particular, we found

that conduction in SiCN ceramics is via the “free” carbon and SiC phases in series, which is

represented by two semicircles in the Nyquist plot. Conduction in SiBCN ceramics is

dominated by one phase, and is thus represented by one semicircle in the Nyquist plot. Such

difference between the SiCN and SiBCN ceramics is due to the presence of nonconducting

BN formed at the interfacial region between the “free” carbon and SiC phases, and isolates

the discontinuous phase from the continuous phase. Therefore, conduction is dominated by

the continuous phase, whether it is “free carbon” or SiC.

4.5.2 Carbon-rich SiCN ceramics as anode material for lithium-ion batteries

This study is focused on the electrochemical investigation of four SiCN ceramics

containing a high content of “free” carbon for use as anode materials for lithium-ion batteries.

The studied SiCN materials are derived from pre-ceramic polymers with different molecular

structures and degrees of branching. For this purpose, branched and linear

polysilylcarbodiimides, as well as branched and linear polysilazanes were synthesized. As

demonstrated by solid-state NMR, the pyrolysis of polysilazanes at 1100°C leads to mixed

SiCxN4-x(x=0-4) bonding configurations within the amorphous SiCN ceramic, as discussed in

Section 4.4. In contrast, the pyrolysis of polysilylcarbodiimides results in the formation of

amorphous nanocomposites comprised of Si3N4, SiC and a “free” carbon phase. This work

focuses on the relationship of the molecular structure of the precursor and the resulting

derived microstructures on the electrochemical properties of the ceramics.


Figure 77 shows the discharge capacities of the investigated materials. The best

performance, with respect to the recovered capacities, was achieved with polysilazane-derived

materials, SiCN3-1100 and SiCN6-1100. Both samples recover the outstanding capacity of

700 mA hg-1 for low current charge/discharge, clearly exceeding the theoretical capacity of

graphite. Starting with a rate of C/5, the capacity of SiCN6 tends to fade, while it remains

perfectly stable for SiCN3-1100. At C/1, the average discharge capacities are 437 mAhg-1 and

316 mAhg-1 for SiCN3-1100 and SiCN6-1100, respectively. The capacities recovered at low

current for the polysilylcarbodiimide-derived samples are much lower, namely 612 and 486

mAhg-1 for the SiCN1-1100 and SiCN5-1100 ceramic, respectively. The SiCN5 ceramic

demonstrates high irreversible capacity during the first cycle, followed by stable

electrochemical cycling behavior. In contrast, SiCN1-1100 shows lower irreversible capacity

during the first cycle, but significant capacity fading at current rates of C/20, C/10 and C/5.

When finally charged/discharged at rates of C/2 and C/1, the capacity stabilizes, and the

recovered capacity is similar to that of SiCN5-1100. The high irreversible capacity of SiCN5-

1100 during the first cycle can be attributed to its high oxygen content.

0 20 40 60 80 100 120 1400

100

200

300

400

500

600

700

800

C/10C/5

C/1C/2

C/20

SiCN1SiCN3SiCN5SiCN6 theor. cap. Graphite

Dis

char

ge

Cap

acit

y [m

A h

g-1]

Cycle Number

C/20

0

100

200

300

400

Cu

rren

t [m

A g

-1]

Figure 77 Discharge capacities with corresponding rates/currents for SiCN1, SiCN3, SiCN5 and SiCN6 prepared

at 1100C. Reprinted from Journal of Power Sources, V236, L. M. Reinold, M. Graczyk-Zajac, Y. Gao, G. Mera, and R. Riedel, Carbon-rich SiCN ceramics as high capacity/high stability anode material for lithium-ion batteries, 224, copyright (2013), with permission from Elsevier [137].

A summary of the corresponding specific capacity values for the 1st and 134th cycles

are given in Table 14. The irreversible capacity Cirreversible represents the amount of charge

which is not recovered during the first discharging process, and the Coulombic efficiency, ,


is given by the ratio (Creversible/Ccharge) x 100%. The efficiency, , was introduced in order to

evaluate the cycling stability of the material, and is given by the ratio of the reversible

capacity recovered after 134 cycles to that after 1 cycle, (Creversible134th/Creversible1st) × 100%.

The averaged discharge capacities were calculated for each current step separately. All

investigated ceramics exhibit promising results in terms of capacity. Our electrochemical

studies clearly show no dependence of the capacity on the “free” carbon amount, and no

dependence on the N/O ratio. These results are clearly in contrast to the hypothesis that a

nitrogen to oxygen ratio smaller than 1 is necessary for SiC(O)N PDCs to achieve suitable

capacities [129]. According to

Figure 77, there is no difference in the discharge capacity of the sample with the

highest N/O ratio (SiCN5) and the sample with the lowest N/O ratio (SiCN1). Furthermore,

the cycling stability of SiCN5-1100 is much better than that of sample SiCN1-1100.

Nevertheless, it should be pointed out that the samples investigated in Ref. [129] contain less

“free” carbon (up to 26 wt%) than the samples investigated in this work.

Table 14 First cycle charging/discharging capacity, irreversible capacity and coulombic efficiency, as well as averaged discharge capacities for different rates.

Sample Ccharge [mA hg-1]

Creversible [mA hg-1]

Cirreversible [mA hg-1] [%]

C/20_start [mA hg-1]

C/10 [mA hg-1]

C/5 [mA hg-1]

C/2 [mA hg-1]

C/1 [mA hg-1]

C/20_end [mA hg-1] [%]

SiCN1-1100 1033 612 421 59 584 483 365 270 235 368 61

SiCN3-1100 1078 703 375 65 699 653 596 516 437 618 89

SiCN5-1100 930 486 444 52 474 397 343 281 240 420 89

SiCN6-1100 1211 724 487 60 715 659 559 430 316 582 82

4.5.3 Summary

The AC conductivity of PDCs studied by impedance spectroscopy reveals

microstructure-based conduction mechanisms. The conduction in SiCN ceramics is via the

“free” carbon and SiC phases in series, which is represented by two semicircles in the

Nyquist plot. Conduction in SiBCN ceramics is dominated by one phase, and is thus

represented by one semicircle in the Nyquist plot. Nonconducting BN forms the interfacial

phase between the “free” carbon and SiC phases, and isolates the discontinuous phase from

the continuous phase. Therefore, conduction in SiBCN ceramics is dominated by the

continuous phase, whether it is “free carbon” or SiC.

The electrochemical study of four SiCN ceramics reveals their outstanding behavior as

high capacity/high stability anode materials for lithium-ion batteries. A linear polysilazane-

derived SiCN ceramic shows excellent performance in terms of cycling stability and capacity

even at high currents. Furthermore, the results demonstrate that a high N/O ratio is not

deleterious to the electrochemical performance of SiC(O)N ceramics. The use of different

classes of precursors does not only lead to different microstructures within the ceramics, but


also to differences in the electrochemical behavior. These investigations nicely bridge

nanostructural features with macroscopic properties.

Conclusion 117

5. Conclusion

In this dissertation, carbon rich SiCN and SiBCN PDCs were produced from tailored

polymers, including linear polysilylcarbodiimides and polysilazanes, their boron-modified

counterparts, as well as branched polysilylcarbodiimide and polysilazane, aiming at obtaining

ceramics with systematically varied compositions and microstructures. For synthesizing

polysilazanes, the salt-free reflux route was employed using hexamethyldisilazane and

dichlorosilane as the reactants and pyridine as the catalyst. NMR spectroscopic analysis

confirmed the molecular structure of the polymers. Both powder and bulk ceramics were

prepared. Powder samples were synthesized by directly pyrolyzing the precursors, while bulk

samples were produced via warm-pressing of the polymer powder prior to pyrolysis. For

thermoset polysilylcarbodiimides and polysilazanes, bulk ceramics were obtained by warm-

pressing with optimized pre-crosslinking parameters. For thermoplastic polyborosilazanes,

bulk samples were obtained using a modified warm-pressing with pre-pyrolyzed materials as

self-fillers. These ceramics were then applied to investigate the characteristic features of

carbon rich PDCs. The following conclusions can be drawn due to our studies:

1. Effects of processing route on crystallization and decomposition of PDCs

The studies of phenyl-containing poly(boro)silylcarbodiimides derived Si(B)CN

powder and bulk ceramics show that the thermal stability against crystallization of the

ceramics strongly depends on the processing route and the chemistry of the precursors. In

SiCN and SiBCN powder samples prepared at 1400 ºC, SiC crystallites are embeded in a

matrix of amorphous carbon and Si3N4. In contrast, the corresponding bulk ceramics remain

completely amorphous at this temperature. The presence of boron in the powder SiBCN

samples promotes SiC crystallization. Moreover, annealing the same powder and bulk

samples up to 2100 oC, HT-TG/DTA showed that bulk ceramics exhibit less weight loss than

their powder analogues, especially for SiCN. The larger surface area of the powder samples

enhances the carbothermal reaction and the crystallization of silicon carbide. The boron

modification was found to impede the degradation of silicon nitride.

2. Effects of structure on thermodynamic stability of PDCs

The oxide melt solution calorimetry study indicates that the ceramic derived from

linear polyphenylvinylsilylcarbodiimide is energetically less stable than that of the ceramic

derived from a branched polyphenylsilsesquicarbodiimide. Multinuclear NMR spectroscopy

demonstrated that the structure of SiCN ceramics derived from polyphenylvinylsilylcarbo-

diimide consists predominantly of separate amorphous domains of silicon nitride and sp2

carbon, while that derived from polyphenylsilsesquicarbodiimide has mixed bonds between N,

C, and Si atoms at the interface between the Si3N4 and C nanodomains. The energetic

stabilization of the latter is then attributed to the presence of such mixed bonds, the presence

of which is likely due to the nature of the branched polymer used in the synthesis of these

PDCs. The relative concentrations of these mixed bonds decrease upon increase of the

118 Conclusion

pyrolysis temperature, along with concomitant hydrogen loss. It is suggested that the mixed

bonde are stabilized via hydrogen bonds located at the interfacial regions between the Si3N4

and C nanodomains, and that the loss of hydrogen at higher temperature provides the entropic

driving force for the destruction of the mixed bonds. The effect of pyrolysis temperature on

the structural evolution of amorphous poly(boro)silazane-derived Si(B)CN was investigated

as well. By MAS NMR spectroscopy, the main processes governing the structural evolution

between 1100 and 1400°C are identified as (i) demixing of SiCxN4-x (x=0-4) mixed bonds, (ii)

cleavage of mixed bonds at interdomain regions, and (iii) coarsening of the domain.

Calorimetry studies demonstrate that this structural evolution is favorable in both enthalpy

and free energy.

3. The relationship between final structure and precursor chemistry

The structure of the resultant ceramics greatly depends on the chemistry of the

corresponding precursors. The results derived from MAS NMR spectroscopy support that the

carbon rich amorphous Si(B)CN ceramics derived from poly(boro)silylcarbodiimides contain

separate Si3N4 and sp2 carbon domains, while ceramics derived from poly(boro)silazanes

consist of mixed bonded SiCxN4-x (x=0-4) tetrahedra. The SiCxN4-x (x=0-4) domains consist

of a core of SiN4 tetrahedra connected to “free” carbon domains via SiN3C, SiN2C2, SiNC3,

and SiC4 tetrahedra. SiC4 tetrahedra make up the most outer shell of the SiCxN4-x(x=0-4)

nanodomains. The SiBCN ceramics derived from boron-modified precursors are comprised of

an additional boron-containing phase which connects the Si-containing domain with that of

the “free” carbon domain. SAXS results indicate that (i) for all precursors, SiCxN4-x (x=0-4)

domain sizes are smaller in samples pyrolyzed at 1100 ºC than at 1400 ºC; (ii) boron-modified

ceramics have smaller SiCxN4-x (x=0-4) domains than their boron-free analogues; (iii)

ceramics derived from poly(boro)silylcarbodiimides have smaller SiCxN4-x (x=0-4) domains

than those derived from poly(boro)silazanes; and (iv) generally, lower carbon contents

contribute to the formation of larger SiCxN4-x (x=0-4) domains.

4. The AC conductivity of different phases reflecting structural information

The AC conductivity of PDCs studied by impedance spectroscopy reveals

microstructure-based conduction mechanisms. Conduction in SiCN ceramics is dominated via

“free” carbon and SiC phases in series, as analyzed by two semicircles in the Nyquist plot.

Conduction in SiBCN ceramics is dominated by one phase, and is thus represented by one

semicircle in the Nyquist plot. Nonconducting BN forms the interfacial phase between the

“free” carbon and SiC phases, and isolates the discontinuous phase from the continuous phase.

Therefore, conduction in SiBCN ceramics is dominated by the continuous phase, whether it is

“free carbon” or SiC.

5. Electrochemical behavior for lithium-ion battery anodes

The electrochemical study of four SiCN ceramics reveals their outstanding behavior as

high capacity/high stability anode materials for lithium-ion batteries. A linear polysilazane-

Conclusion 119

derived SiCN ceramic shows excellent performance in terms of cycling stability and capacity,

even at high currents. Furthermore, the results demonstrate that a high N/O ratio is not

deleterious to the electrochemical performance of SiC(O)N ceramics. The use of different

classes of precursors leads to different SiCN microstructures, which in turn explains their

significant differences in electrochemical behavior.

120 Outlook

6. Outlook

This thesis presents a systematic study of the nanodomain structures in SiCN and

SiBCN PDCs synthesized from tailored polymers. New routes for polymer synthesis were

explored, and their optimized bulk processing contributes valuably to the PDC field. The

characterization techniques used for studying the microstructure, solid state structure, and

energetic stability of these materials provide some guidelines for future studies. Detailed

structural information of the ceramic nanodomains derived from the synthesized precursors

can be used as reference for future studies, particularly since many common characteristics

exist in the microstructure of PDCs derived from similar precursor systems.

Although this study was systematic and complete for its own purpose, additional

investigations are necessary to further understand in more detail the nature of the preceramic

polymers, the derived ceramics, and the polymer-to-ceramic transformation. An outline of

possible investigations is recommended as follows:

1. As discussed, three types of reactions occur during pyrolysis: (i) demixing of

SiCxN4-x (x=0-4) tetrahedra, (ii) diminishing of mixed bonds at interdomain regions, and (iii)

coarsening of nanodomains. Calorimetry gives the total energetic consequence of these

processes through the change in enthalpy. However, except for the obvious notion that

nanodomain coarsening is energetically stabilizing, the energetic effects of the other processes

are unclear. Whether these effects are stabilizing or destabilizing, and to what extent, remains

unknown. By carrying out similar study involving materials synthesized at more intermediate

pyrolysis temperatures, the structural evolution could be probed more thoroughly, in order to

determine the onset of bond demixing and phase separation. Coupling these results with

calorimetry studies will provide more insight into the nature of the bond demixing process.

2. The high frequency response semicircles obtained in the impedance spectroscopy

investigations are not completed due to instrumental limitations. Measurements up to at least

several GHz are necessary in order to obtain more detailed information. Along with improved

structural characterizations of the materials, the AC conductivity contributed by each phase

should be better assessed.

References 121

References

[1] P. Colombo, G. Mera, R. Riedel, and G. D. Sorarù, J. Am. Ceram. Soc. 93, 1805 (2010).

[2] G. Mera, A. Tamayo, H. Nguyen, S. Sen, and R. Riedel, J. Am. Ceram. Soc. 93, 1169 (2010).

[3] A. Saha, R. Raj, D. L. Williamson, and H.-J. Kleebe, J. Am. Ceram. Soc. 88, 232 (2005).

[4] A. Saha, R. Raj, and D. L. Williamson, J. Am. Ceram. Soc. 89, 2188 (2006).

[5] G. Mera, R. Riedel, F. Poli, and K. Müller, J. Eur. Ceram. Soc. 29, 2873 (2009).

[6] Y. Iwamoto, W. Völger, E. Kroke, R. Riedel, T. Saitou, and K. Matsunaga, J. Am. Ceram. Soc. 84, 2170 (2001).

[7] H. J. Seifert, H. L. Lukas, and F. Aldinger, Berichte Bunsenges. Für Phys. Chem. 102, 1309 (1998).

[8] M. Weinmann, J. Schuhmacher, H. Kummer, S. Prinz, J. Peng, H. J. Seifert, M. Christ, K. Müller, J. Bill, and F. Aldinger, Chem. Mater. 12, 623 (2000).

[9] H. J. Seifert, J. Peng, J. Golczewski, and F. Aldinger, Appl. Organomet. Chem. 15, 794 (2001).

[10] R. M. Morcos, G. Mera, A. Navrotsky, T. Varga, R. Riedel, F. Poli, and K. Müller, J. Am. Ceram. Soc. 91, 3349 (2008).

[11] R. M. Morcos, A. Navrotsky, T. Varga, D. Ahn, A. Saha, F. Poli, K. Müller, and R. Raj, J. Am. Ceram. Soc. 91, 2391 (2008).

[12] R. M. Morcos, A. Navrotsky, T. Varga, Y. Blum, D. Ahn, F. Poli, K. Müller, and R. Raj, J. Am. Ceram. Soc. 91, 2969 (2008).

[13] A. H. Tavakoli, J. A. Golczewski, J. Bill, and A. Navrotsky, Acta Mater. 60, 4514 (2012).

[14] T. Varga, A. Navrotsky, J. L. Moats, R. M. Morcos, F. Poli, K. Müller, A. Saha, and R. Raj, J. Am. Ceram. Soc. 90, 3213 (2007).

[15] A. H. Tavakoli, P. Gerstel, J. A. Golczewski, and J. Bill, Acta Mater. 58, 6002 (2010).

[16] A. H. Tavakoli, P. Gerstel, J. A. Golczewski, and J. Bill, J. Mater. Res. 25, 2150 (2011).

[17] Z.-C. Wang, F. Aldinger, and R. Riedel, J. Am. Ceram. Soc. 84, 2179 (2001).

[18] P. Gerstel, A. Müller, J. Bill, and F. Aldinger, Chem. Mater. 15, 4980 (2003).

[19] Z.-C. Wang, P. Gerstel, G. Kaiser, J. Bill, and F. Aldinger, J. Am. Ceram. Soc. 88, 2709 (2005).

[20] R. Riedel, G. Mera, R. Hauser, and A. Klonczynski, J. Ceram. Soc. Jpn. 114, 425 (2006).

[21] R. Riedel, G. Passing, H. Schönfelder, and R. J. Brook, Nature 355, 714 (1992).

[22] G. Passing, H. Schönfelder, R. Riedel, and R. Brook, Br Ceram Trans 92, 21 (1993).

[23] J. Seitz and J. Bill, J. Mater. Sci. Lett. 15, 391 (1996).

[24] R. Haug, M. Weinmann, J. Bill, and F. Aldinger, J. Eur. Ceram. Soc. 19, 1 (1999).

[25] Konetschny, C., J. Eur. Ceram. Soc. 19, 2789 (1999).

[26] J. Wan, M. J. Gasch, and A. K. Mukherjee, J. Mater. Res. 15, 1657 (2000).

122 References

[27] J. Seitz, B. Joachim, F. Aldinger, and Y. Naerheim, U.S. Patent No. 6458315 (1 October 2002).

[28] S. R. Shah and R. Raj, Acta Mater. 50, 4093 (2002).

[29] R. Harshe, C. Balan, and R. Riedel, J. Eur. Ceram. Soc. 24, 3471 (2004).

[30] N. Janakiraman and F. Aldinger, J. Eur. Ceram. Soc. 29, 163 (2009).

[31] T. Jiang, A. Hill, W. Fei, Y. Wei, M. Tellam, C. Xu, and L. An, J. Am. Ceram. Soc. 93, 3017 (2010).

[32] D. Su, Y.-L. Li, H.-J. An, X. Liu, F. Hou, J.-Y. Li, and X. Fu, J. Eur. Ceram. Soc. 30, 1503 (2010).

[33] N. Janakiraman, M. Weinmann, J. Schuhmacher, K. Müller, J. Bill, F. Aldinger, and P. Singh, J. Am. Ceram. Soc. 85, 1807 (2002).

[34] N. Janakiraman, A. Zern, M. Weinmann, F. Aldinger, and P. Singh, J. Eur. Ceram. Soc. 25, 509 (2005).

[35] E. Barsoukov and J. R. Macdonald, Impedance Spectroscopy, Theory Experiment and Applications (John Wiley & Sons, Inc., 2005).

[36] C. Haluschka, C. Engel, and R. Riedel, J. Eur. Ceram. Soc. 20, 1365 (2000).

[37] S. Trassl, M. Puchinger, E. Rössler, and G. Ziegler, J. Eur. Ceram. Soc. 23, 781 (2003).

[38] D. Ahn and R. Raj, J. Power Sources 196, 2179 (2011).

[39] R. Ralf, M. Gabriela, H. Ralf, and K. Alexander, J. Ceram. Soc. Jpn. 114, 425 (2006).

[40] P. Colombo, Polymer Derived Ceramics: From Nano-Structure to Applications (DEStech Publications, Inc, 2010).

[41] B. Papendorf, K. Nonnenmacher, E. Ionescu, H.-J. Kleebe, and R. Riedel, Small 7, 970 (2011).

[42] R. Riedel, A. Kienzle, W. Dressler, L. Ruwisch, J. Bill, and F. Aldinger, Nature 382, 796 (1996).

[43] A. O. Gabriel, R. Riedel, S. Storck, and W. F. Maier, Appl. Organomet. Chem. 11, 833 (1997).

[44] A. O. Gabriel, R. Riedel, W. Dressler, S. Reichert, C. Gervais, J. Maquet, and F. Babonneau, Chem. Mater. 11, 412 (1999).

[45] R. Raj, R. Riedel, and G. D. Soraru, J. Am. Ceram. Soc. 84, 2158 (2001).

[46] R. Riedel, E. Kroke, A. Greiner, A. O. Gabriel, L. Ruwisch, J. Nicolich, and P. Kroll, Chem. Mater. 10, 2964 (1998).

[47] R. Riedel and A. O. Gabriel, Adv. Mater. 11, 207 (1999).

[48] J. Schuhmacher, M. Weinmann, J. Bill, F. Aldinger, and K. Müller, Chem. Mater. 10, 3913 (1998).

[49] J. E. Drake, R. T. Hemmings, and E. Henderson, J. Chem. Soc. Dalton Trans. 366 (1976).

[50] J. E. Drake, B. M. Glavinčevski, H. E. Henderson, and R. T. Hemmings, Synth. React. Inorg. Met.-Org. Chem. 8, 7 (1978).

[51] P. Haag, R. Lechler, and J. Weidlein, Z. Für Anorg. Allg. Chem. 620, 112 (1994).

[52] J. J. Moscony and A. G. MacDiarmid, Chem. Commun. Lond. 307 (1965).

[53] R. Riedel, A. Greiner, G. Miehe, W. Dressler, H. Fuess, J. Bill, and F. Aldinger, Angew. Chem. Int. Ed. Engl. 36, 603 (1997).

References 123

[54] S. Nahar-Borchert, E. Kroke, R. Riedel, B. Boury, and R. J. P. Corriu, J. Organomet. Chem. 686, 127 (2003).

[55] S. Nahar-Borchert, A. O. Gabriel, and R. Riedel, in Mater. Synth., edited by U. Schubert, N. Hüsing, and R. M. Laine (Springer Vienna, 2004), pp. 193–198.

[56] M. Peuckert, T. Vaahs, and M. Brück, Adv. Mater. 2, 398 (1990).

[57] D. Bahloul, M. Pereira, P. Goursat, N. S. C. K. Yive, and R. J. P. Corriu, J. Am. Ceram. Soc. 76, 1156 (1993).

[58] Y. D. Blum, K. B. Schwartz, and R. M. Laine, J. Mater. Sci. 24, 1707 (1989).

[59] R. J. P. C. N. S. Choong Kwet Yive, New J Chem 15, 85 (1991).

[60] E. Kroke, Y.-L. Li, C. Konetschny, E. Lecomte, C. Fasel, and R. Riedel, Mater. Sci. Eng. R Reports 26, 97 (2000).

[61] A. Lavedrine, D. Bahloul, P. Goursat, N. Choong Kwet Yive, R. Corriu, D. Leclerq, H. Mutin, and A. Vioux, J. Eur. Ceram. Soc. 8, 221 (1991).

[62] D. Seyferth, G. H. Wiseman, and C. Prud’homme, J. Am. Ceram. Soc. 66, C (1983).

[63] D. Seyferth, G. H. Wiseman, and C. Prud’homme, in Emergent Process Methods High-Technol. Ceram., edited by R. F. Davis, H. P. III, and R. L. Porter (Springer US, 1984), pp. 263–269.

[64] T. Vaahs, M. Brück, and W. D. G. Böcker, Adv. Mater. 4, 224 (1992).

[65] E. Werner, U. Klingebiel, F. Pauer, D. Stalke, R. Riedel, and S. Schaible, Z. Für Anorg. Allg. Chem. 596, 35 (1991).

[66] W. NOLL, in Chem. Technol. Silicones (Academic Press, 1968).

[67] Y. D. Blum and R. M. Laine, U.S. Patent No. 4801439 (31 January 1989).

[68] Y. Blum and R. M. Laine, Organometallics 5, 2081 (1986).

[69] C. Xu, N. Zhou, and Z. Xie, Chin. Chem. Lett. 11, 765 (2000).

[70] J. P. Mooser, H. Nöth, and W. Tinhof, Z Naturforsch 29b, 166 (1974).

[71] M. Weinmann, R. Haug, J. Bill, M. De Guire, and F. Aldinger, Appl. Organomet. Chem. 12, 725 (1998).

[72] K. B. Schwartz and D. j. Rowcliffe, J. Am. Ceram. Soc. 69, C (1986).

[73] S. Zhu, W. G. Fahrenholtz, and G. E. Hilmas, Scr. Mater. 59, 123 (2008).

[74] S. T. Schwab, C. R. Blanchard, and R. C. Graef, J. Mater. Sci. 29, 6320 (1994).

[75] M. Balog, J. Keckes, T. Schöberl, D. Galusek, F. Hofer, J. Krest’ An, Z. Lences, J.-L. Huang, and P. Sajgalik, J. Eur. Ceram. Soc. 27, 2145 (2007).

[76] O. Funayama, T. Kato, Y. Tashiro, and T. Isoda, J. Am. Ceram. Soc. 76, 717 (1993).

[77] D. Seyferth and H. Plenio, J. Am. Ceram. Soc. 73, 2131 (1990).

[78] K. Su, L. G. Sneddon, E. E. Remsen, and G. A. Zank, Chem. Mater. United States 5:4, (1993).

[79] T. Wideman, E. Cortez, E. E. Remsen, G. A. Zank, P. J. Carroll, and L. G. Sneddon, Chem. Mater. 9, 2218 (1997).

[80] M. Takamizawa, T. Kobayashi, A. Hayashida, and Y. Takeda, U.S. Patent No. 4604367 (5 August 1986).

[81] R. Riedel and M. Seher, J. Eur. Ceram. Soc. 7, 21 (1991).

[82] L. L. Snead and S. J. Zinkle, Nucl. Instruments Methods Phys. Res. Sect. B Beam Interactions Mater. Atoms 191, 497 (2002).

[83] F. Aldinger, M. Weinmann, and J. Bill, Pure Appl. Chem. 70, 439 (1998).

124 References

[84] H.-P. Baldus, G. Passing, H. Scholz, D. Sporn, M. Jansen, and J. Göring, Key Eng. Mater. 127-131, 177 (1997).

[85] S. Bernard, M. Weinmann, D. Cornu, P. Miele, and F. Aldinger, J. Eur. Ceram. Soc. 25, 251 (2005).

[86] J. Bill and F. Aldinger, Adv. Mater. 7, 775 (1995).

[87] M. Christ, G. Thurn, M. Weinmann, J. Bill, and F. Aldinger, J. Am. Ceram. Soc. 83, 3025 (2000).

[88] M. Christ and A. Zimmermann, J. Mater. Res. 16, 1994 (2001).

[89] A. Müller, P. Gerstel, M. Weinmann, J. Bill, and F. Aldinger, J. Eur. Ceram. Soc. 2655 (2000).

[90] R. Riedel, L. M. Ruswisch, L. An, and R. Raj, J. Am. Ceram. Soc. 81, 3341 (1998).

[91] M. Weinmann, T. W. Kamphowe, J. Schuhmacher, K. Müller, and F. Aldinger, Chem. Mater. 12, 2112 (2000).

[92] A. Navrotsky, Phys. Chem. Miner. 2, 89 (1977).

[93] A. Navrotsky, Phys. Chem. Miner. 24, 222 (1997).

[94] Y. Zhang, A. Navrotsky, and T. Sekine, J Mater Res 21, 41 (2006).

[95] J. Liang, L. Topor, A. Navrotsky, and M. Mitomo, J Mater Res 14, 1959 (1999).

[96] P. Kroll, J. Non-Cryst. Solids 351, 1121 (2005).

[97] A. H. Tavakoli, P. Gerstel, J. A. Golczewski, and J. Bill, J. Mater. Res. 26, 600 (2011).

[98] G. Mera, R. Riedel, F. Poli, and K. Müller, J. Eur. Ceram. Soc. 29, 2873 (2009).

[99] J. Seitz, J. Bill, N. Egger, and F. Aldinger, J. Eur. Ceram. Soc. 16, 885 (1996).

[100] S. Traßl, G. Motz, E. Rössler, and G. Ziegler, J. Am. Ceram. Soc. 85, 239 (2002).

[101] S. Traßl, H.-J. Kleebe, H. Störmer, G. Motz, E. Rössler, and G. Ziegler, J. Am. Ceram. Soc. 85, 1268 (2002).

[102] S. Traßl, D. Suttor, G. Motz, E. Rössler, and G. Ziegler, J. Eur. Ceram. Soc. 20, 215 (2000).

[103] S. Traßl, G. Motz, E. Rössler, and G. Ziegler, J. Non-Cryst. Solids 293–295, 261 (2001).

[104] J. Schuhmacher, F. Berger, M. Weinmann, J. Bill, F. Aldinger, and K. Müller, Appl. Organomet. Chem. 15, 809 (2001).

[105] J. Bill, T. W. Kamphowe, A. Müller, T. Wichmann, A. Zern, A. Jalowieki, J. Mayer, M. Weinmann, J. Schuhmacher, K. Müller, J. Peng, H. J. Seifert, and F. Aldinger, Appl. Organomet. Chem. 15, 777 (2001).

[106] F. Berger, A. Müller, F. Aldinger, and K. Müller, Z. Für Anorg. Allg. Chem. 631, 355 (2005).

[107] W. Gruber, O. Starykov, W. Oppermann, and H. Schmidt, Diffus. Fundam. 8, 9 (2008).

[108] W. Gruber, O. Starykov, and H. Schmidt, Phys. Status Solidi RRL - Rapid Res. Lett. 3, 85 (2009).

[109] J. Haug, P. Lamparter, M. Weinmann, and F. Aldinger, Chem. Mater. 16, 72 (2004).

[110] J. Haug, P. Lamparter, M. Weinmann, and F. Aldinger, Chem. Mater. 16, 83 (2004).

[111] N. Resta, C. Kohler, and H.-R. Trebin, J. Am. Ceram. Soc. 86, 1409 (2003).

[112] J. Cordelair and P. Greil, J. Eur. Ceram. Soc. 20, 1947 (2000).

[113] C. Engel, Elektrische Eigenschaften gefüllter Pyrolyse-Keramiken im System Si-O-C (Shaker, 2000).

References 125

[114] S. Trassl, G. Motz, E. Rössler, and G. Ziegler, J. Non-Cryst. Solids 293–295, 261 (2001).

[115] S. Trassl, H.-J. Kleebe, H. Störmer, G. Motz, E. Rössler, and G. Ziegler, J. Am. Ceram. Soc. 85, 1268 (2002).

[116] J. Robertson, Adv. Phys. 35, 317 (1986).

[117] S. Wansom, N. J. Kidner, L. Y. Woo, and T. O. Mason, Cem. Concr. Compos. 28, 509 (2006).

[118] H. Fukui, H. Ohsuka, T. Hino, and K. Kanamura, ACS Appl. Mater. Interfaces 2, 998 (2010).

[119] M. Graczyk-Zajac, L. Toma, C. Fasel, and R. Riedel, Solid State Ionics 225, 522 (2012).

[120] J. Kaspar, M. Graczyk-Zajac, and R. Riedel, Solid State Ionics 225, 527 (2012).

[121] A. M. Wilson, J. N. Reimers, E. W. Fuller, and J. R. Dahn, Solid State Ionics 74, 249 (1994).

[122] J. R. Dahn, A. M. Wilson, W. Xing, and G. A. Zank, U.S. Patent No. US5631106 A (20 May 1997).

[123] D. Su, Y.-L. Li, Y. Feng, and J. Jin, J. Am. Ceram. Soc. 92, 2962 (2009).

[124] Y. Feng, Electrochimica Acta 55, 5860 (2010).

[125] J. Kaspar, G. Mera, A. P. Nowak, M. Graczyk-Zajac, and R. Riedel, Electrochimica Acta 56, 174 (2010).

[126] M. Graczyk-Zajac, G. Mera, J. Kaspar, and R. Riedel, J. Eur. Ceram. Soc. 30, 3235 (2010).

[127] M. Graczyk-Zajac, C. Fasel, and R. Riedel, J. Power Sources 196, 6412 (2011).

[128] L. M. Reinold, M. Graczyk-Zajac, C. Fasel, and R. Riedel, ECS Trans. 35, 37 (2011).

[129] D. Ahn and R. Raj, J. Power Sources 196, 2179 (2011).

[130] I. A. Vostokov and Y. I. Dergunov, Zh Obshch Khim 47, 1769 (1977).

[131] S. Widgeon, G. Mera, Y. Gao, S. Sen, A. Navrotsky, and R. Riedel, J. Am. Ceram. Soc. 96, 1651 (2013).

[132] J. Liang, A. Navrotsky, V. J. Leppert, M. J. Paskowitz, S. H. Risbud, T. Ludwig, H. J. Seifert, F. Aldinger, and M. Mitomo, J. Mater. Res. 14, 4630 (1999).

[133] H. Nguyen, Synthese und Charakterisierung von polymerabgeleiteten Si(B)CN Keramiken, Diplomarbeit, Technische Universität Darmstadt, 2010.

[134] Y. Gao, G. Mera, H. Nguyen, K. Morita, H.-J. Kleebe, and R. Riedel, J. Eur. Ceram. Soc. 32, 1857 (2012).

[135] Y. Gao, S. J. Widgeon, T. B. Tran, A. H. Tavakoli, G. Mera, S. Sen, R. Riedel, and A. Navrotsky, Scr. Mater. 69, 347 (2013).

[136] S. Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel, Chem. Mater. 24, 1181 (2012).

[137] L. M. Reinold, M. Graczyk-Zajac, Y. Gao, G. Mera, and R. Riedel, J. Power Sources 236, 224 (2013).

[138] A. C. Ferrari and J. Robertson, Phys. Rev. B 61, 14095 (2000).

[139] M. A. Pimenta, G. Dresselhaus, M. S. Dresselhaus, L. G. Cançado, A. Jorio, and R. Saito, Phys. Chem. Chem. Phys. 9, 1276 (2007).

[140] A. C. Ferrari, J. C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. S. Novoselov, S. Roth, and A. K. Geim, Phys. Rev. Lett. 97, (2006).

126 References

[141] A. C. Ferrari, Solid State Commun. 143, 47 (2007).

[142] L. G. Cançado, K. Takai, T. Enoki, M. Endo, Y. A. Kim, H. Mizusaki, A. Jorio, L. N. Coelho, R. Magalhães-Paniago, and M. A. Pimenta, Appl. Phys. Lett. 88, 163106 (2006).

[143] A. Sadezky, H. Muckenhuber, H. Grothe, R. Niessner, and U. Pöschl, Carbon 43, 1731 (2005).

[144] G. A. Zickler, B. Smarsly, N. Gierlinger, H. Peterlik, and O. Paris, Carbon 44, 3239 (2006).

[145] R. Saito, A. Jorio, A. G. Souza Filho, G. Dresselhaus, M. S. Dresselhaus, and M. A. Pimenta, Phys. Rev. Lett. 88, 027401 (2002).

[146] M. M. Lucchese, F. Stavale, E. H. M. Ferreira, C. Vilani, M. V. O. Moutinho, R. B. Capaz, C. A. Achete, and A. Jorio, Carbon 48, 1592 (2010).

[147] E. H. Martins Ferreira, M. V. O. Moutinho, F. Stavale, M. M. Lucchese, R. B. Capaz, C. A. Achete, and A. Jorio, Phys. Rev. B 82, 125429 (2010).

[148] A. Jorio, E. H. Martins Ferreira, L. G. Cançado, C. A. Achete, and R. B. Capaz, in Phys. Appl. Graphene - Exp., edited by S. Mikhailov (InTech, 2011).

[149] H.-J. Kleebe, Phys. Status Solidi 166, 297 (1998).

[150] C. Turquat, H.-J. Kleebe, G. Gregori, S. Walter, and G. D. Sorarù, J. Am. Ceram. Soc. 84, 2189 (2001).

[151] H. Bréquel, J. Parmentier, S. Walter, R. Badheka, G. Trimmel, S. Masse, J. Latournerie, P. Dempsey, C. Turquat, A. Desmartin-Chomel, L. Le Neindre-Prum, U. A. Jayasooriya, D. Hourlier, H.-J. Kleebe, G. D. Sorarù, S. Enzo, and F. Babonneau, Chem. Mater. 16, 2585 (2004).

[152] A. H. Tavakoli, P. Gerstel, J. A. Golczewski, and J. Bill, J. Non-Cryst. Solids 355, 2381 (2009).

[153] S. Martinez-Crespiera, E. Ionescu, H.-J. Kleebe, and R. Riedel, J. Eur. Ceram. Soc. 31, 913 (2011).

[154] A. Müller, P. Gerstel, M. Weinmann, J. Bill, and F. Aldinger, J. Eur. Ceram. Soc. 21, 2171 (2001).

[155] A. H. Tavakoli, Thermodynamic and Kinetic Studies on the Thermal Stability of Amorphous Si-(B-)C-N Ceramics, 2010.

[156] J. Dürr, S. Schempp, P. Lamparter, J. Bill, S. Steeb, and F. Aldinger, Solid State Ionics 101, 1041 (1997).

[157] A. Saha, R. Raj, and D. L. Williamson, J. Am. Ceram. Soc. 060428035142017 (2006).

[158] E. Leonova, J. Grins, M. Shariatgorji, L. L. Ilag, and M. Edén, Solid State Nucl. Magn. Reson. 36, 11 (2009).

[159] E. Kolehmainen, K. Laihia, J. Korvola, V. S. Kaganovich, M. I. Rybinskaya, and Z. A. Kerzina, J. Organomet. Chem. 487, 215 (1995).

[160] V. A. Du, G. N. Stipicic, and U. Schubert, Eur. J. Inorg. Chem. 2011, 3365 (2011).

[161] A. Scarmi, G. D. Sorarù, and R. Raj, J. Non-Cryst. Solids 351, 2238 (2005).

[162] Y. H. Sehlleier, A. Verhoeven, and M. Jansen, J. Mater. Chem. 17, 4316 (2007).

[163] W. J. Gammon, D. I. Malyarenko, O. Kraft, G. L. Hoatson, A. C. Reilly, and B. C. Holloway, Phys. Rev. B 66, 153402 (2002).

[164] H. Imai, P. K. Babu, E. Oldfield, A. Wieckowski, D. Kasuya, T. Azami, Y. Shimakawa, M. Yudasaka, Y. Kubo, and S. Iijima, Phys. Rev. B 73, 125405 (2006).

References 127

[165] C. Gervais, F. Babonneau, L. Ruwisch, R. Hauser, and R. Riedel, Can. J. Chem. 81, 1359 (2003).

[166] C. W. Bale, P. Chartrand, S. A. Degterov, G. Eriksson, K. Hack, R. Ben Mahfoud, J. Melançon, A. D. Pelton, and S. Petersen, Calphad 26, 189 (2002).

[167] R. A. Robie and B. S. Hemingway, Thermodynamic Properties of Minerals and Related Substances at 295.15K and 1 Bar (105 Pa) Pressure and at Higher Temperatures (1995).

[168] S. Widgeon, G. Mera, Y. Gao, S. Sen, A. Navrotsky, and R. Riedel, J Am Ceram Soc (2013).

[169] G. Beaucage, S. Rane, D. W. Schaefer, G. Long, and D. Fischer, J. Polym. Sci. Part B Polym. Phys. 37, 1105 (1999).

[170] N. Larouche and B. L. Stansfield, Carbon 48, 620 (2010).

[171] J. R. Macdonald, Impedance Spectroscopy: Emphasizing Solid Materials and Systems (Wiley, 1987).

[172] K. Wang, B. Ma, Y. Wang, and L. An, J. Am. Ceram. Soc. 96, 1363 (2013).

128 Curriculum Vitae

Curriculum Vitae

Yan GAO

Address： Rabenaustr. 9

64293, Darmstadt

Germany

E-mail：[email protected]

Personal status

Gender: Female

Date of Birth: 04/24/1981

Citizenship: China

Scientific background

1999-2003 Bachelor in Materials Science & Engineering, Tsinghua University, P.R.

China

Thesis: Gel-casting of fused silica ceramic

Advisor: Prof. Zhipeng Xie

2003-2006 Master in Materials Science & Engineering, Tsinghua University, P.R.

China

Thesis: Injection molding of silicon carbide and zirconia and the removal of water-soluble

binders

Advisor: Prof. Zhipeng Xie

2010-2013 Ph.D. in Material Science, Technische Universität Darmstadt (TUD),

Germany

Thesis: Nanodomain structure and energetics of carbon rich SiCN and SiBCN polymer-

derived ceramics

Advisor: Prof. Ralf Riedel

Research experiences

2010-2013 Research Assistant, Universität Darmstadt

Synthesized the preceramic polymers with tailored compositions and structures, and

the isotope counterparts.

Developed the processing technology that can make compact bulk ceramics derived

from polymers.

Studied the structural evolution of polymer-derived ceramics depending on their

unique polymer chemistry. Investigated the influenc of the nanostructures of polymer-

Curriculum Vitae 129

derived amorphous ceramics on their thermodynamic stabilities.

Investigated the electrical and electrochemical behavior of amorphous ceramics

derived from different precursors. Developed fundamental understandings of these

properties.

03-04/2011 and 08-09/2012 Research Assistant, University of California, Davis

Investigated on the thermal stability of amorphous PDCs using oxidative drop solution

calorimetry and high-temperature TGA.

Studied nanodomain structures of amorphous PDCs with solid state MAS NMR.

2003-2006 Research Assistant, Tsinghua University

Developed a new organic system based on partially water-soluble binders and

associated debinding technique for injection molding of ceramics. The efficiency of

the debinding process was greatly improved compared to the traditional process using

wax based binders. Optimized the process of injection molding. Improved the

mechanical properties and micro-structures of molded parts.

Developed a novel processing technique for making fine and/or bio-compatible

ceramic components/parts, such as dental posts, dental brackets, optical fiber sleeves,

fan shafts and razors. The technique has been transferred to industries and the

fabricated components/parts have been put into applications in hospitals and industries.

2003 Research Assistant, Tsinghua University

Studied and improved colloidal system for Gel-casting process applications.

Improved dispersibility of the colloidal suspension by optimizing process parameters,

such as pH value, particle size and its distribution, etc. Fabricated amorphous fused

silica ceramic with good mechanical properties.

Working experience

2006-2010 Engineer, China Aerospace Science and Technology Corporation

Carried out the investigation on the development and potential use of composites.

Studied the functionality of space materials, complying with the safe applications in

the extreme space environment, such as high/low temperature, thermal cycling,

vacuum and radiation, etc.

Served as a technical support for qualification approval of the manufacturers of space

used composite materials.

2005 Internship, Foxconn Technological Corporation

Developed a new organic system for injection molding, which reduced the

manudacturing cost.

130 Curriculum Vitae

Fabricated zirconia optical fiber connectors, reduced the defects of products.

Activities

Teaching the lab courses of “Polymer-derived SiOC ceramic” and “Silicon Nitride

ceramic” in 2011/2012/2013 at TU Darmstadt.

Supervise the “School girls” program in 2010/2011/2012, at TU Darmstadt

Publications

Ph.D.

1) Y. Gao, G. Mera, H. Nguyen, K. Morita, H-J. Kleebe, and R. Riedel. Processing route

dramatically influencing the nanostructure of carbon-rich SiCN and SiBCN polymer-derived

ceramics. Part I: Low temperature thermal transformation J. Eur. Ceram. Soc., 32[9], 1857-

1866 (2012).

2) Y. Gao, S. Widgeon, T. Tran, A. H. Tavakoli, G. Mera, S. Sen, R. Riedel, and A.

Navrotsky. Effect of Demixing and Coarsening on the Energetics of Poly(boro)silazane-

Derived Amorphous Si-(B-)C-N Ceramics, Scr. Mater. 69, 347(2013).

3) S. Widgeon, G. Mera, Y. Gao, E. Stoyanov, S. Sen, A. Navrotsky, and R. Riedel.

Nanostructure and Energetics of Carbon-Rich SiCN Ceramics Derived from

Polysilylcarbodiimides: Role of the Nanodomain Interfaces. Chem. Mater., 24, 1181 (2012).

4) S. Widgeon, G. Mera, Y. Gao, S. Sen, A. Navrotsky, and R. Riedel. Effect of precursor

on speciation and nanostructure of SiBCN polymer derived ceramics. J. Am. Ceram. Soc., 96,

1651 (2013).

5) L. M. Reinold, M. Graczyk-Zajac, Y. Gao, G. Mera, and R. Riedel. Carbon-rich SiCN

ceramics as high capacity/high stability anode material for lithium ion batteries. Journal of

Power Sources, 236, 224–229 (2013).

6) Y. Gao, M. Bazarjani, S. Sen, R. Riedel. Nanodomain Characteristic in Polymer-Derived

Si(B)CN Ceramics. To be submitted.

7) Y. Gao, H. Nguyen, Y. Xu, G. Buntkowsky and R. Riedel. Development of

Thermoplastic Polyborosilazanes for Polymer-Derived SiBCN Ceramic Monoliths. In

preparation.

Master & Bachelor

8) Y. Gao, K. Huang, Z. Fan, and Z. Xie. Water-soluble binder based Injection molding of

zirconia. Key Eng. Mater., 336-338, 1017-1020, 2007.

9) Y. Gao, Z. Xie, and Y. Huang. Gel-casting of fused silica ceramic and its sintering

process. Rare Metal Mater. and Eng., 34, 438-442, 2005.

Curriculum Vitae 131

10) J. Luo, Z. Yi, B. Xiao, Y. Gao, Z. Xie, J. Li, and Y. Huang. Injection molding of ultra-

fine zirconia(Y-TZP) powders. J. Ceram. Proc. Tech., 7[1], 14-19, 2006.

11) Z. Xie, B. Xiao, Y. Gao, J. Li, and Y. Huang. Injection molding of high performance

ceramic parts. Adv. Manuf. Mater. Appl. Tech., 4, 13-16, 2005.

Conference Presentation

12) Y. Gao, G. Mera, H. Nguyen, K. Morita, H-J. Kleebe, and R. Riedel. Processing of Bulk

Carbon-Rich SiCN and SiBCN Ceramics Derived from Polysilylcarbodiimides and

Polyborosilylcarbodiimides Presentation in “the seventh Intenational Conference on High-

performance Ceramics”, Xiamen, China, Nov. 4-7, 2011.

Other Publications:

13) As Editor, English-Chinese Dictionary of Composite Materials Technology, ISBN 7-

5025-9295-4, Chemical Industry Press, Jan. 2007, the First Edition.

14) Y. Gao, Plan for the aerospace composite materials standards, Standards and Civilian

Aerospace, 2007.

15) Y. Gao et al. Space systems - safety and compatibility of materials - Part 1-Part 5:

Aerospace Standards of China.

16) Y. Gao, Aerospace composites materials assurance practice, Aerospace Standardization,

2009(4), pp. 43~46.

17) Z. Gao and Y. Gao, Evaluation requirements for qualified composite material supplier of

space products, Aerospace Standards of China.

18) H. Zhang, Y. Zhu, J. Yan, Z. Gao, and Y. Gao. The classification and code of space

materials, Aerospace Standards of China.

Awards & Honors

2000 Third Class Kang-hong Scholarship for Excellent Students of Tsinghua University.

2005 Third Class Scholarship for Excellent Integrated Graduate Students of Tsinghua

University.

2005 Excellent League Member of Tsinghua University.

2007 Outstanding Employee of CASI

2008 Advanced Individual of CASI

2009 Second Class Technical Achievement of CASI

132 Eidesstattliche Erklärung

Eidesstattliche Erklärung

M. Sc. Yan Gao

Rabenaustr. 9

64293 Darmstadt

Eidesstattliche Erklärung

Hiermit erkläre ich an Eides statt, dass ich die vorliegende Dissertation selbstständig und nur mit den angegebenen Hilfsmitteln angefertigt habe. Von mir wurde weder an der Technischen Universität Darmstadt noch einer anderen Hochschule ein Promotionsversuch unternommen.

Darmstadt, den 09.10.2013

Yan Gao

Documents

tuprints.ulb.tu-darmstadt.de Gao... · A Dissertation approved by the department of Material Science In fulillment of the requirements for the degree of Doktor−Ingenieur (Dr.‐Ing.)